Filtering options for reflected signal with noise

>> The signal we are going to be using will be somewhere around 500-1kHz,
>> 20V. The main component of noise is going to be centered around 50Hz
>> (England) and then several harmonics above this.
>
>Depending on the impedance, several hundred amperes might be needed to 
>attain that voltage. You really need some preliminary measurements or a 
>confirmed model in order to settle on such decisions.
>
>> If I were to connect to the rails with an isolation transformer this
would
>> remove the common-mode noise (most of the 50Hz and harmonics).
>
>An isolation transformer is, in theory, sensitive to differential-mode 
>voltages only. In practice, parasitic capacitance may give it weak 
>sensitivity to common-mode voltage, but "weak" is the operative word.
>
>> Would the isolation transformer have any negative impact upon the
>> differential-mode signal when it comes to measuring them for impedance
>> calculations.
>
>No.
>
>> My section cannot effect the next section. All I ultimately
>> have to do is get the signal off the lines and remove any noise which
is
>> present, before the readings can occur.
>
>If sections are isolated, what completes the traction-current circuit?

I meant my filtering section/subsystem cannot effect the readings which
are going to be taken.

There is no break in the lines so the traction current can continue
along.
Track circuits do use impedance bonds which only allows the traction
current through and filters the signal.

I was thinking that to ensure successive devices do not effect each other,
each would operate on a slightly different frequency. I could then place a
bandpass active filter after the transformer which filters the frequency
for that particular section.

There would only need to be 3-4 different frequencies before they can be
repeated.

>> The signal we are going to be using will be somewhere around 500-1kHz,
>> 20V. The main component of noise is going to be centered around 50Hz
>> (England) and then several harmonics above this.
>
>Depending on the impedance, several hundred amperes might be needed to 
>attain that voltage. You really need some preliminary measurements or a 
>confirmed model in order to settle on such decisions.
>
>> If I were to connect to the rails with an isolation transformer this
would
>> remove the common-mode noise (most of the 50Hz and harmonics).
>
>An isolation transformer is, in theory, sensitive to differential-mode 
>voltages only. In practice, parasitic capacitance may give it weak 
>sensitivity to common-mode voltage, but "weak" is the operative word.
>
>> Would the isolation transformer have any negative impact upon the
>> differential-mode signal when it comes to measuring them for impedance
>> calculations.
>
>No.
>
>> My section cannot effect the next section. All I ultimately
>> have to do is get the signal off the lines and remove any noise which
is
>> present, before the readings can occur.
>
>If sections are isolated, what completes the traction-current circuit?

I meant my filtering section/subsystem cannot effect the readings which
are going to be taken.

There is no break in the lines so the traction current can continue
along.
Track circuits do use impedance bonds which only allows the traction
current through and filters the signal.

I was thinking that to ensure successive devices do not effect each other,
each would operate on a slightly different frequency. I could then place a
bandpass active filter after the transformer which filters the frequency
for that particular section.

There would only need to be 3-4 different frequencies before they can be
repeated.

John121 wrote:
>>> The signal we are going to be using will be somewhere around 500-1kHz,
>>> 20V. The main component of noise is going to be centered around 50Hz
>>> (England) and then several harmonics above this.
>> Depending on the impedance, several hundred amperes might be needed to 
>> attain that voltage. You really need some preliminary measurements or a 
>> confirmed model in order to settle on such decisions.
>>
>>> If I were to connect to the rails with an isolation transformer this
> would
>>> remove the common-mode noise (most of the 50Hz and harmonics).
>> An isolation transformer is, in theory, sensitive to differential-mode 
>> voltages only. In practice, parasitic capacitance may give it weak 
>> sensitivity to common-mode voltage, but "weak" is the operative word.
>>
>>> Would the isolation transformer have any negative impact upon the
>>> differential-mode signal when it comes to measuring them for impedance
>>> calculations.
>> No.
>>
>>> My section cannot effect the next section. All I ultimately
>>> have to do is get the signal off the lines and remove any noise which
> is
>>> present, before the readings can occur.
>> If sections are isolated, what completes the traction-current circuit?
> 
> I meant my filtering section/subsystem cannot effect the readings which
> are going to be taken.
> 
> There is no break in the lines so the traction current can continue
> along.
> Track circuits do use impedance bonds which only allows the traction
> current through and filters the signal.

Impedance bonds? Do you mean those pieces of cable that are welded 
across rail joints bypassing the fish plates? Are they somehow frequency 
selective in a useful way?

> I was thinking that to ensure successive devices do not effect each other,
> each would operate on a slightly different frequency. I could then place a
> bandpass active filter after the transformer which filters the frequency
> for that particular section.

That makes sense, but it requires narrow-band selective filters. Noise 
filters in addition don't make a sensible part of that picture.

> There would only need to be 3-4 different frequencies before they can be
> repeated.

Good. How do you know?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

John121 wrote:
>>> The signal we are going to be using will be somewhere around 500-1kHz,
>>> 20V. The main component of noise is going to be centered around 50Hz
>>> (England) and then several harmonics above this.
>> Depending on the impedance, several hundred amperes might be needed to 
>> attain that voltage. You really need some preliminary measurements or a 
>> confirmed model in order to settle on such decisions.
>>
>>> If I were to connect to the rails with an isolation transformer this
> would
>>> remove the common-mode noise (most of the 50Hz and harmonics).
>> An isolation transformer is, in theory, sensitive to differential-mode 
>> voltages only. In practice, parasitic capacitance may give it weak 
>> sensitivity to common-mode voltage, but "weak" is the operative word.
>>
>>> Would the isolation transformer have any negative impact upon the
>>> differential-mode signal when it comes to measuring them for impedance
>>> calculations.
>> No.
>>
>>> My section cannot effect the next section. All I ultimately
>>> have to do is get the signal off the lines and remove any noise which
> is
>>> present, before the readings can occur.
>> If sections are isolated, what completes the traction-current circuit?
> 
> I meant my filtering section/subsystem cannot effect the readings which
> are going to be taken.
> 
> There is no break in the lines so the traction current can continue
> along.
> Track circuits do use impedance bonds which only allows the traction
> current through and filters the signal.

Impedance bonds? Do you mean those pieces of cable that are welded 
across rail joints bypassing the fish plates? Are they somehow frequency 
selective in a useful way?

> I was thinking that to ensure successive devices do not effect each other,
> each would operate on a slightly different frequency. I could then place a
> bandpass active filter after the transformer which filters the frequency
> for that particular section.

That makes sense, but it requires narrow-band selective filters. Noise 
filters in addition don't make a sensible part of that picture.

> There would only need to be 3-4 different frequencies before they can be
> repeated.

Good. How do you know?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

John121 wrote:
>>> The signal we are going to be using will be somewhere around 500-1kHz,
>>> 20V. The main component of noise is going to be centered around 50Hz
>>> (England) and then several harmonics above this.
>> Depending on the impedance, several hundred amperes might be needed to 
>> attain that voltage. You really need some preliminary measurements or a 
>> confirmed model in order to settle on such decisions.
>>
>>> If I were to connect to the rails with an isolation transformer this
> would
>>> remove the common-mode noise (most of the 50Hz and harmonics).
>> An isolation transformer is, in theory, sensitive to differential-mode 
>> voltages only. In practice, parasitic capacitance may give it weak 
>> sensitivity to common-mode voltage, but "weak" is the operative word.
>>
>>> Would the isolation transformer have any negative impact upon the
>>> differential-mode signal when it comes to measuring them for impedance
>>> calculations.
>> No.
>>
>>> My section cannot effect the next section. All I ultimately
>>> have to do is get the signal off the lines and remove any noise which
> is
>>> present, before the readings can occur.
>> If sections are isolated, what completes the traction-current circuit?
> 
> I meant my filtering section/subsystem cannot effect the readings which
> are going to be taken.
> 
> There is no break in the lines so the traction current can continue
> along.
> Track circuits do use impedance bonds which only allows the traction
> current through and filters the signal.

Impedance bonds? Do you mean those pieces of cable that are welded 
across rail joints bypassing the fish plates? Are they somehow frequency 
selective in a useful way?

> I was thinking that to ensure successive devices do not effect each other,
> each would operate on a slightly different frequency. I could then place a
> bandpass active filter after the transformer which filters the frequency
> for that particular section.

That makes sense, but it requires narrow-band selective filters. Noise 
filters in addition don't make a sensible part of that picture.

> There would only need to be 3-4 different frequencies before they can be
> repeated.

Good. How do you know?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

John121 wrote:
>>> The signal we are going to be using will be somewhere around 500-1kHz,
>>> 20V. The main component of noise is going to be centered around 50Hz
>>> (England) and then several harmonics above this.
>> Depending on the impedance, several hundred amperes might be needed to 
>> attain that voltage. You really need some preliminary measurements or a 
>> confirmed model in order to settle on such decisions.
>>
>>> If I were to connect to the rails with an isolation transformer this
> would
>>> remove the common-mode noise (most of the 50Hz and harmonics).
>> An isolation transformer is, in theory, sensitive to differential-mode 
>> voltages only. In practice, parasitic capacitance may give it weak 
>> sensitivity to common-mode voltage, but "weak" is the operative word.
>>
>>> Would the isolation transformer have any negative impact upon the
>>> differential-mode signal when it comes to measuring them for impedance
>>> calculations.
>> No.
>>
>>> My section cannot effect the next section. All I ultimately
>>> have to do is get the signal off the lines and remove any noise which
> is
>>> present, before the readings can occur.
>> If sections are isolated, what completes the traction-current circuit?
> 
> I meant my filtering section/subsystem cannot effect the readings which
> are going to be taken.
> 
> There is no break in the lines so the traction current can continue
> along.
> Track circuits do use impedance bonds which only allows the traction
> current through and filters the signal.

Impedance bonds? Do you mean those pieces of cable that are welded 
across rail joints bypassing the fish plates? Are they somehow frequency 
selective in a useful way?

> I was thinking that to ensure successive devices do not effect each other,
> each would operate on a slightly different frequency. I could then place a
> bandpass active filter after the transformer which filters the frequency
> for that particular section.

That makes sense, but it requires narrow-band selective filters. Noise 
filters in addition don't make a sensible part of that picture.

> There would only need to be 3-4 different frequencies before they can be
> repeated.

Good. How do you know?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

John121 wrote:
>>> The signal we are going to be using will be somewhere around 500-1kHz,
>>> 20V. The main component of noise is going to be centered around 50Hz
>>> (England) and then several harmonics above this.
>> Depending on the impedance, several hundred amperes might be needed to 
>> attain that voltage. You really need some preliminary measurements or a 
>> confirmed model in order to settle on such decisions.
>>
>>> If I were to connect to the rails with an isolation transformer this
> would
>>> remove the common-mode noise (most of the 50Hz and harmonics).
>> An isolation transformer is, in theory, sensitive to differential-mode 
>> voltages only. In practice, parasitic capacitance may give it weak 
>> sensitivity to common-mode voltage, but "weak" is the operative word.
>>
>>> Would the isolation transformer have any negative impact upon the
>>> differential-mode signal when it comes to measuring them for impedance
>>> calculations.
>> No.
>>
>>> My section cannot effect the next section. All I ultimately
>>> have to do is get the signal off the lines and remove any noise which
> is
>>> present, before the readings can occur.
>> If sections are isolated, what completes the traction-current circuit?
> 
> I meant my filtering section/subsystem cannot effect the readings which
> are going to be taken.
> 
> There is no break in the lines so the traction current can continue
> along.
> Track circuits do use impedance bonds which only allows the traction
> current through and filters the signal.

Impedance bonds? Do you mean those pieces of cable that are welded 
across rail joints bypassing the fish plates? Are they somehow frequency 
selective in a useful way?

> I was thinking that to ensure successive devices do not effect each other,
> each would operate on a slightly different frequency. I could then place a
> bandpass active filter after the transformer which filters the frequency
> for that particular section.

That makes sense, but it requires narrow-band selective filters. Noise 
filters in addition don't make a sensible part of that picture.

> There would only need to be 3-4 different frequencies before they can be
> repeated.

Good. How do you know?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

John121 wrote:
>>> The signal we are going to be using will be somewhere around 500-1kHz,
>>> 20V. The main component of noise is going to be centered around 50Hz
>>> (England) and then several harmonics above this.
>> Depending on the impedance, several hundred amperes might be needed to 
>> attain that voltage. You really need some preliminary measurements or a 
>> confirmed model in order to settle on such decisions.
>>
>>> If I were to connect to the rails with an isolation transformer this
> would
>>> remove the common-mode noise (most of the 50Hz and harmonics).
>> An isolation transformer is, in theory, sensitive to differential-mode 
>> voltages only. In practice, parasitic capacitance may give it weak 
>> sensitivity to common-mode voltage, but "weak" is the operative word.
>>
>>> Would the isolation transformer have any negative impact upon the
>>> differential-mode signal when it comes to measuring them for impedance
>>> calculations.
>> No.
>>
>>> My section cannot effect the next section. All I ultimately
>>> have to do is get the signal off the lines and remove any noise which
> is
>>> present, before the readings can occur.
>> If sections are isolated, what completes the traction-current circuit?
> 
> I meant my filtering section/subsystem cannot effect the readings which
> are going to be taken.
> 
> There is no break in the lines so the traction current can continue
> along.
> Track circuits do use impedance bonds which only allows the traction
> current through and filters the signal.

Impedance bonds? Do you mean those pieces of cable that are welded 
across rail joints bypassing the fish plates? Are they somehow frequency 
selective in a useful way?

> I was thinking that to ensure successive devices do not effect each other,
> each would operate on a slightly different frequency. I could then place a
> bandpass active filter after the transformer which filters the frequency
> for that particular section.

That makes sense, but it requires narrow-band selective filters. Noise 
filters in addition don't make a sensible part of that picture.

> There would only need to be 3-4 different frequencies before they can be
> repeated.

Good. How do you know?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

>Impedance bonds? Do you mean those pieces of cable that are welded 
>across rail joints bypassing the fish plates? Are they somehow frequency

>selective in a useful way?

Well I thought impedance bonds allowed the traction current to pass and
attenuated the signal frequency of the particular track circuit.

>That makes sense, but it requires narrow-band selective filters. Noise 
>filters in addition don't make a sensible part of that picture.

Could you please explain this slightly more? What do you mean by noise
filters in addition?

>> There would only need to be 3-4 different frequencies before they can
be
>> repeated.

Well I know that sending a 1kHz, 20V frequency attenuates to less than 1V
after 1 mile, therefore over 3 miles it should be small enough to not cause
any noticeable change to be used again.

John121 wrote:
>> Impedance bonds? Do you mean those pieces of cable that are welded 
>> across rail joints bypassing the fish plates? Are they somehow frequency
> 
>> selective in a useful way?
> 
> Well I thought impedance bonds allowed the traction current to pass and
> attenuated the signal frequency of the particular track circuit.

I think they're short pieces heavy of heavy wire. Do we have the same 
things in mind?

>> That makes sense, but it requires narrow-band selective filters. Noise 
>> filters in addition don't make a sensible part of that picture.
> 
> Could you please explain this slightly more? What do you mean by noise
> filters in addition?

Presumably, your noise filters would pass the interesting signal and 
block the noise that occurs at other frequencies. The tuning (channel 
selection) filters would do the same thing.

>>> There would only need to be 3-4 different frequencies before they can
> be
>>> repeated.
> 
> Well I know that sending a 1kHz, 20V frequency attenuates to less than 1V
> after 1 mile, therefore over 3 miles it should be small enough to not cause
> any noticeable change to be used again.

I hope you're right. I guess you are, but it needs to be confirmed by test.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������