RE: Signal vs. Idle debate (A picture is worth a thousand words)
Mr Polk,
I take offense to your comments; this type of flaming response is not the type
of e-mail that we promote on this reflector. Your e-mail was inciting,
unprofessional, unconstructive, and hardly conducive to the type interaction we
foster in the IEEE for communications over the reflector. You sir, owe me an
apology.
I obviously touched on what must be a sensitive issue for you. I will address
your comments one by one.
See comments below.
...Ralph
"James M. Polk" <jmpolk@xxxxxxxxx> on 05/05/2000 06:53:03 PM
To: Ralph Andersson/TSC/TDK-US@TSC, "DOVE,DANIEL J
(HP-Roseville,ex1)" <dan_dove@xxxxxx>
cc: stds-802-3-pwrviamdi@xxxxxxxx
Subject: RE: Signal vs. Idle debate (A picture is worth a
thousand words)
>Ralph
>At 04:21 PM 5/5/2000 -0700, Ralph.Andersson@xxxxxxxxxxx wrote:
>>Phantom power via the signal pair requires the addition of an additional
>>transformer in the signal path; the effect on the return loss is unknown and
is
>>may be detrimental.
>So, in your expert opinion.... and in your words BTW, the phrase "... the
>effect on the return loss is unknown and is may be detrimental." means we
>shouldn't even try, huh....
>Is that considered scientific.... or logical.... or even deductive decision
>making?
This comment is two-fold:
First, to date, there has been no presentation on the effects on return loss
when an additional transformer is added to the data link. Look for this to
change at the Ottawa Interim. My company, TDK Semiconductor, voted against the
motion to use the idle pair at the March Plenary. We thought the vote premature
as the data was not present to support such a vote. We at TDK are maintaining
an open mind. Are you?
Second, I have direct knowledge of how tight the return loss spec is for
100BASE-TX data links. It may not be easy to design a transformer that will
solve this problem. Again, we need data to determine the magnitude of the
problem. I guess a company can always choose to ignore the return loss spec.
As an engineer, I wouldn't do this.
There is another effect that must be considered along with the return loss
issue. There is an additional pole added by each transformer inserted into the
data link. How many legacy PHYs might stop working by the addition of another
transformer? Again, the data has not been presented to make such a decision.
>Well how many thousands of years, in how many cultures, did the idea of flying
>not get pursued because it *may* fail.
>Where's the innovative engineering effort here?
>I don't know if you're an engineer -- but how about trying... and failing
>before you shout at the top of your lungs something won't work. That is the
>scientific and logical way.
Please re-read my e-mail. Actually it was presented in an ambivalent manner.
As pointed out in the previously included comment from Geoff, he seems to share
the concern over lack of margin for the insertion of new elements. ...and he
wasn't shouting either.
>I know our products work with power applied to the signaling pairs -- and we
>have end station devices that take that power without hesitation at 100BASE-Tx
>communications speeds (which seems to be in question from some -- I even think
>we demonstrated that with 2 other companies at the January meeting. I don't
>know if we're the only company with such a solution -- but I'd be surprised if
>no one else couldn't engineer what we have.
A solution that works is not necessarily the best one to implement. Until all
of the data is in we do not want to implement changes that might adversely
affect a large base of installed equipment. We are doing the due diligence to
make sure that the published, final solution is the best.
>So... IMO, if there is NO MARKET for powering down the signaling pairs, I guess
>we should just stop trying.... it just isn't worth the effort, it can't happen,
>can it?
See above. Again, my comments were fairly ambivalent w.r.t. what the solution
should be. The important point is that the 10BASE-T and 100BASE-TX solutions
should not be jeopardized because one company adopted a solution that uses the
signal pairs. With this in mind we should look for the data that supports the
addition of a transformer in the signal pair. I haven't seen this data.
>Is my sarcasm constructive? No.
>I know it isn't.
>I'm just surprised most of this list won't even try this option. Every one of
>the emails I've read on this topic keep saying *may* and *unknown effect*....
>therefore chose the *other (admittedly easier) way*.....
If you have data that conclusively addresses the concerns I (and others) have
raised I will be happy to review it for you.
>*************************************
>"At the end of the day... the most committed win!"
>James M. Polk
>Sr. Product Manager, Multiservice Architecture and Standards
>Enterprise Voice Business Unit
>Cisco Systems
>Dallas, Texas
>w) 972.813.5208
>f) 972.813.5280
>www.cisco.com