Methods and devices for detecting electromagnetic interference on data transmission lines
View Patent ↗A method for remotely detecting electromagnetic interference on a data transmission line of a communication network includes measuring at a client site a noise power spectrum on the data transmission line as a function of frequency, for a frequency range. The noise is present when the data transmission line is not carrying a data signal. The method further includes forwarding the measured noise power spectrum to a remote computer at a network operator's site, determining periodicities in the measured power spectrum by the remote computer, and determining the switching frequency of a power supply unit at the client site by the remote computer based on periodicities in the power spectrum.
1. A method for remotely detecting electromagnetic interference on a data transmission line of a communication network, in particular for electromagnetic interference caused by a power supply unit located at a client site, said power supply unit including a switching converter having a switching frequency, comprising:
measuring at a client site a power spectrum of a noise signal on said data transmission line as a function of frequency, for a predetermined frequency range, said noise being present when said data transmission line is not carrying a data signal;
forwarding said measured noise power spectrum to a remote computing means at a network operator's site;
determining periodicities in said measured power spectrum by said remote computing means;
determining the switching frequency of said power supply unit by said remote computing means, based on periodicities in said power spectrum.
2. A method according to claim 1 , comprising:
determining a YIN estimator, said YIN estimator being defined as
d
′
(
Δ
f
)
=
{
1
,
if
Δ
f
=
0
d
(
τ
)
1
Δ
f
∑
j
=
1
Δ
f
d
(
j
)
,
otherwise
with
d
(
Δ
f
)
=
∑
j
=
1
N
(
x
j
-
x
j
+
Δ
f
)
2
wherein 11f is a frequency difference value, xj is the power spectrum value of the signal for discrete signal frequencies j, and N is the number of discrete signal frequencies within said predetermined width of the power spectrum of said noise signal;
determining the minimum value for said YIN estimator;
associating the minimum value of said YIN estimator with the switching frequency of said power supply unit.
3. A method according to claim 1 , wherein determining periodicities in said power spectrum comprises:
determining an autocorrelation function for said measured power spectrum;
determining the maximum of said autocorrelation function;
associating said switching frequency of said power supply unit with the maximum of said autocorrelation function.
4. A method according to claim 1 , wherein determining periodicities in said power spectrum comprises:
determining a cepstrum of said noise signal, said cepstrum being the square of the Fourier transform of the 10-based logarithm of the squared magnitude of the fourier transform of said noise signal;
deriving a frequency value corresponding to the largest peak value for said cepstrum;
associating said switching frequency of said power supply unit with said frequency value.
5. A method according to claim 1 , wherein determining periodicities in said power spectrum comprises:
detecting a set of peak values and associated peak frequencies in said power spectrum;
determining a set of difference values, said values consisting of the differences between each pair of frequencies of said set of peak frequencies;
counting the number of occurrences of similar difference values for said set of difference values;
associating said switching frequency of said power supply unit with the difference value which has the largest number of occurrences.
6. A method according to claim 1 , wherein said predetermined frequency range comprises the whole power spectrum of said noise signal.
7. A method according to claim 1 , further comprising:
detecting which frequency range of signals on said line is impacted by said power supply unit, by performing the method according to claim 2 a plurality of times for a plurality of different frequency ranges of a predetermined width, said predetermined width being smaller than the width of said whole noise spectrum.
8. A method according to claim 1 , further comprising:
detecting which frequency range of signals on said line is impacted by said power supply unit, including,
determining a two-dimensional Yin estimator defined as;
d
′
(
Δ
f
,
f
0
)
=
{
1
,
if
Δ
f
=
0
d
(
Δ
f
,
f
0
)
1
Δ
f
∑
j
=
1
N
d
(
j
,
f
0
)
,
otherwise
with
d
(
Δ
f
,
f
0
)
=
∑
j
=
f
0
f
0
+
W
(
x
j
-
x
j
+
Δ
f
)
2
wherein 11f is a frequency difference value, fo a start frequency, Xj is the power spectrum value of the signal for discrete signal frequencies J, and W is width of the part of the signal for which the YIN estimator is computed, this width being smaller than the whole power spectrum of said noise signal;
determining a two-dimensional minimum value area of said two-dimensional estimator; and
associating said two-dimensional minimum value area with the switching frequency of said power supply unit and with said a frequency range which is impacted by said power supply unit.
9. A method according to claim 1 , further comprising:
detecting which frequency range of signals on said line is impacted by said power supply unit, including,
determining a two-dimensional Yin estimator defined as;
d
′
(
Δ
f
,
f
0
)
=
{
1
,
if
Δ
f
=
0
d
(
Δ
f
,
f
0
)
1
Δ
f
∑
j
=
1
N
d
(
j
,
f
0
)
,
otherwise
with
d
(
Δ
f
,
f
0
)
=
∑
j
=
f
0
f
0
+
k
Δ
f
(
x
j
-
x
j
+
Δ
f
)
2
wherein 11f is a frequency difference value, fo a start frequency, Xj is the power spectrum value of the signal for discrete signal frequencies j, and k is a predetermined multiplication factor;
determining a two-dimensional minimum value area of said two-dimensional estimator;
associating said two-dimensional minimum value area with the switching frequency of said power supply unit and with said a frequency range which is impacted by said power supply unit.
10. A method according to claim 1 , further comprising:
looking up said determined switching frequency in a predetermined information list, said list including information relating to predetermined types of power supply units and their associated switching frequencies, thereby identifying said type of said power supply unit.
11. A method according to claim 10 , further comprising:
forwarding information regarding said type of power supply unit by said remote computing means to at least one client device on said client site.
12. A method according to claim 10 , further comprising:
storing said switching frequencies for use in network management purposes.
13. A remote computing means for detecting electromagnetic interference on a data transmission line of a communication network, in particular for electromagnetic interference caused by a power supply unit located at a client site, said power supply unit comprising:
a switching converter, including,
a reception means for receiving power spectrum measurements for a noise signal on a transmission line as a function of frequency for a predetermined frequency range from said client site, said noise signal being present when said transmission line is not carrying a data signal; and
a calculation means for determining periodicities in said power spectrum measurements and for determining the switching frequency of said power supply unit, based on periodicities in said power spectrum.
14. A remote computing means according to claim 13 , further comprising:
a means for looking up said determined switching frequency in a predetermined information list, said list including information relating to predetermined types of power supply units and their associated switching frequencies, thereby identifying said type of said power supply unit.
15. A remote computing means according to claim 13 , further comprising:
a means for forwarding information regarding said type of power supply unit to at least one client device.