RE: [10GBASE-T] channel model
Albert,
The impulse response corresponding to the proposed channel model has been verified by measurement at room temperature.
Since the temperature correction method affects only magnitude but not phase it will introduce some small amount of phase error into the complex-valued channel model. You are correct that this error will appear in the time domain as a small amount of non-causal impulse response energy. Note the non-causal element disappears if the temperature correction is removed. The added temperature correction (from 20C to 50C) can be easily removed by following the procedure in ASTM D4566 Section 26.4. The resulting "uncorrected" model will be free of non-causal artifacts.
The proposed model was intended as an approximation for a basic feasibility study and a starting point for defining high frequency channel insertion loss (magnitude only). As such the emphasis was on obtaining an accurate frequency domain representation. I chose a specific temperature correction method that was defined in an existing standard, but the method was intended to adjust magnitude only. While the time domain error is an issue for detailed implementation studies, it should not be an issue for the intended applications. In fact, simply blanking out the non-causal portion should provide sufficiently accurate information for a basic implementation study. Please note the proposed model is a starting point. Refinements and corrections will be added in the future as needed.
Regards,
Larry
-----Original Message-----
From: Albert Vareljian [mailto:albertv@ieee.org]
Sent: Wednesday, January 29, 2003 6:15 PM
To: Larry Cohen
Cc: William Jones; stds-802-3-10gbt@ieee.org
Subject: Re: [10GBASE-T] channel model
Larry,
Has the impulse response corresponding to the proposed
channel model frequency characteristic been verified? This
is important, as the system feasibility can only be reliably
ascertained in the time domain.
The method used for the channel loss temperature correction
mentioned in your correspondence could potentially result in
non-causal time domain behavior for the following reason.
The only [relatively] temperature invariant component in the
line phase response is its non-minimum-phase portion - the
linear part of the phase characteristic that corresponds to the
line propagation delay. However, the so called minimum-phase
portion (non-linear), is known to be related to the cable
magnitude response by the Hilbert Transform. Hence, the
magnitude cannot be in general manipulated arbitrary, without
causing an impact on the phase.
(The minimum-phase component being critical for the system
time domain behavior is usually small comparable to its linear
non-minimum-phase counterpart, which makes direct
measurements difficult.)
The channel impulse response generation would be straightforward
(using IFFT) if the frequency step was chosen to be ~488.28 kHz,
as high order interpolation techniques required for smooth fitting
of the current data could prove unreliable.
Regards,
Albert
Larry Cohen wrote:
>Attached is a proposed 100 meter Cat 5e channel model for 10GBaseT study (CHANMOD.TXT). This model is defined by complex-valued (X + jY format) insertion gain in the at 500 kHz intervals from DC to 1 GHz.
>
>The proposed model is derived from measurement (at 20 C) of a 100 meter channel (with Cat 5e patch cords and 4 Cat 5e RJ45 connector interfaces). The baseline 100 meter channel model was temperature corrected to 50 C. (122 F.) using the procedure defined in ASTM D4566 Section 26.4. The above mentioned temperature correction method modifies the magnitude but not the phase of the insertion loss. Experimental measurements confirm that temperature effects on the phase are minimal so the approximation is sufficiently accurate.
>
>The most recent proposed model is an extrapolation of the ISO 11801 Class D channel loss limit. The ISO 11801 Class D limit is designed to accommodate expected worst-case measurements below 100 MHz and consequently includes some additional margin to meet this requirement. Since this limit is only specified below 100 MHz, extrapolation of the limit through smooth curve fitting becomes less accurate as it is extended farther beyond its defined boundaries. The new proposed model is based upon measurements beyond 500 MHz and thus validated to at least 500 MHz (and somewhat beyond 500 MHz). As shown in the attached graphs (CHANMOD.DOC), the insertion loss for the proposed model differs from the ISO 11801 Class D 2002 channel limit by less than 0.2 dB from DC to 50 MHz and less than 0.5 dB from 50 to 100 MHz. While it does not represent an absolute worst-case channel, it represents a reasonable model for a feasibility study simulating a maximum length channel under worst-case temperature conditions.
>
>-Larry
>
>
>
>-----Original Message-----
>From: William Jones
>Sent: Wednesday, January 22, 2003 11:19 PM
>To: stds-802-3-10gbt@ieee.org
>Subject: RE: [10GBASE-T] channel model
>
>
>
>Xiaopeng
>
>How are the calculations coming? Are you using the models Chris sent out last week?
>
>regards
>Bill
>
>-----Original Message-----
>From: William Jones
>Sent: Wednesday, January 15, 2003 4:08 PM
>To: stds-802-3-10gbt@ieee.org
>Subject: [10GBASE-T] channel model(part 2)
>
>
>
>Bill,
>
>Sure. Once the common channel model (I only need the frequency domain
>characteristics) is available, I can begin to work on it.
>
>Regards,
>Xiaopeng
>----------
>
>Xiaopeng
>
>Would you be willing to sign up for technical feasibility based on an
>optimal DFE calculation?
>
>regards
>Bill
>-----------
>
>Bill,
>
>I has no problem to use TIA or ISO based channel and noise model once it is
>standardized. My point is that even using the frequency-domain model we
>can still give a quite good estimation of the practically achievable
>distance at a throughtput of 10Gbps over CAT-5E or CAT-6 cables.
>
>Regards,
>
>Xiaopeng
>-------------
>
>Xiaopeng
>
>We believe, as George discussed in his part of the tutorial in Kauai, that
>with an adequate amount of crosstalk cancellation, the throughput is
>achievable. The question then becomes can this level of cancellation be
>practically achieved, hence, my interest in time domain models.
>
>Until we get the models from the channel modeling Ad Hoc, we could argue
>this point endlessly. Until then, I still believe we should not use DSL
>models, but, rather something from TIA 568b or ISO 11801.
>
>regards
>
>Bill
>
>
>