# Come back to 100Base-TX - possible?

• posted

Bonjour Rich,

In june 2001, that is yesterday, you wrote that, about 100Base-TX:

"The scrambling prevents IDLEs (and any other repetitive data pattern) from concentrating all of the transmitted energy into a narrow frequency band, and hence violating EMI requirements. (Without scrambling, the MLT-3 encoded IDLE would appear as a continuous, 31.25 MHz continuous-wave signal.)"

I agree with that statement.

In the same conditions, with the scrambling, we have now 125 MHz /

2047 = 61 kHz instead of 31.25 MHz continuous-wave signal.

The value 2047 come from the (2^11 - 1) period with the 11 registers PRBS.

Perhaps the crosstalk can be acceptable, and the energy in this frequency (NEXT or FEXT) would be low.

Thanks if you have a comment. Best regards, Michelot

• posted

If it is periodic at 61kHz you want the Fourier transform (or Fourier series) of the signal to see how it is distributed. Energy will be at multiples of 61kHz. With a good scrambler it will be distributed somewhat evenly among the harmonics, or decreasing with higher frequencies.

This is also true. Both inductive and capacitive crosstalk decrease with decreasing frequency. There will still be energy up to 62.5MHz, assuming the appropriate filter to keep higher harmonics out.

-- glen

• posted

Bonsoir Glen,

Yes, it is

For 100Base-TX, the scrambler is well defined in 802.3.

100Base-TX: 11 registers 1000Base-T: 33 registers 10GBase-R: 58 registers

You're right, we don't have to forget it.

About the fact of the "good scrambler", I don't know if we can construct a PRBS with any generator polynomial (e.g. beyond the 11th degree).

Best regards, Michelot

• posted

So all you need is the output of the scrambler with idle as input. 2047 different voltages, and calculate the Fourier series from that. That will give the amplitude and phase of the harmonics from 61kHz to 62.5MHz in 61kHz steps.

-- glen

• posted

Bonjour Glen,

It's the main reason why the scrambling exists: to avoid high frequencies.

Ok, thanks for these comments. We avoid this high energetic frequency (or energetic frequency band centred around 31.25 MHz).

Best regards, Michelot

• posted

In some cases it is also to ensure a sufficient number of transitions.

For DPSK, for example, a continuous stream of zeros might result in no phase changes, making it hard for the receiver to synchronize with the signal. A scrambler pretty much ensures transitions. (If one was really careful one might generate a signal with no phase transitions for a long time.)

-- glen

• posted

Bonjour Glen,

You're right. So it depends on the case. When the scrambling is after the channel coding (code-bits converted in code-groups), normally the code-groups contents sufficient "1". And, generally, a logical "1" gives a transition after the line coding.

It is the case with 100Base-TX, for 4B/5B and MLT-3. In this case, the scrambling is not done to ensure a sufficient number of transitions. This fonction is ensure by the both codings.

It is certainly the case with 1000Base-CX for 8/10B and ????

Thanks, best regards, Michelot

• posted

Bonsoir Michelot,

It seems that it is NRZ (curious for the copper).

In this case, my supposition is wrong, we have a transition for "1" followed by "0", or the contrary, and not just "1" as MLT-3.

Best regards, Michelot

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.