roy@phri.UUCP (Roy Smith) (07/07/89)
We recently installed a Farallon StarController to turn our passive
star PhoneNet network into an active star. Before I hooked any wire up to
the SC, I ran the built-in line quality tests. About half of the branches
showed Poor or Marginal with no wire connected. Farallon tech support said
not to worry about it, the diagnostics are more sensitive than they should
be with no wire connected.
After I hooked up 4 branches, I still got most of the branches
testing Marginal or Poor. This was surprising, because the net worked even
with all 4 branches just tied together on a 66 block. The only branch that
showed Good was one with only about 40' of 22 gauge wire. This didn't
really surprise me since most of those branches have wiring rules
violations up the wazoo (yes, I've got my project for the summer). What
did surprise me is when I put in a new branch with about 400' of 24 gauge
wire (going through, if I've kept count properly, 6 punchdown blocks) and a
single PhoneNet connector on it, and it too tested out Marginal. Surely
that is well within the specs. Everything works fine, but I'm worried
about the diagnostics.
Does anybody have any experience with these things? When the SC
says a line is Marginal or Poor, should I really worry about it as long as
the net seems to be working?
--
Roy Smith, Public Health Research Institute
455 First Avenue, New York, NY 10016
{allegra,philabs,cmcl2,rutgers,hombre}!phri!roy -or- roy@alanine.phri.nyu.edu
"The connector is the network"roy@phri.UUCP (Roy Smith) (07/12/89)
In <3839@phri.UUCP> I complained about my new StarController always saying branches were "Poor" or "Marginal" when they where almost certainly well within specs. In <12673@netnews.upenn.edu> chip@pender.ee.upenn.edu (Charles H. Buchholtz): > I finally called Farallon, who said that they had recently gone to a > new parts supplier, and had softened the hardware specs so that they > could use cheaper parts. They hadn't gotten around to adjusting the > software yet, and so I should consider Marginal to mean Good. As long > as it didn't say Poor, I shouldn't worry. This seemed pretty dubious to me. I called Farallon and got essentially the same story, except that they added that they had some new software (not yet frozen, but a later beta than what I got) which should fix the problem that they are sending to me. They insist that the new SCs with the new-style transformers are just as good as the old ones (the guy on the phone didn't quite know what to say when I suggested that the opposite might be true) and that it's just a minor change in the software. Personally, I am a bit suspicious of this "if it's broken, let's change the diagnostics" routine, but we'll see what happens. Given that the branches which the SC rates as "Poor" are well within the theoretical wiring limits (aan more importantly, they seem to work fine), I'm willing to believe that the fault really is in over-sensitive down-checking by the diagnostic tests. The test, as far as I can tell, seems to be to drive the line with a pulse train at the standard LocalTalk frequency (230 kbps, or something like that) and then time how long it takes for the line to go quiet after the driving pulse train is stopped. It also seems that Farallon recommends a maximum of 3 punch-down blocks in any branch. I can't find where it says that in the SC documentation. At any rate, some of our branches go through as many as 7 blocks (counting the SC distribution block), plus the 42A block on the end. Had I planned the wiring plant, I might not have done it that way, but I didn't have anything to do with it. It's either use exising spare pairs or run my own trunks. Guess which I did. I suppose every block is another impedence mismatch, but what can I do? -- Roy Smith, Public Health Research Institute 455 First Avenue, New York, NY 10016 {allegra,philabs,cmcl2,rutgers,hombre}!phri!roy -or- roy@alanine.phri.nyu.edu "The connector is the network"
josefk@well.UUCP (William B. Douglas) (07/15/89)
I'd like to clear up a few points that have perhaps become a little muddy in
this thread. I _am_ speaking as a representative of Farallon here, so I will
try to be brief.
1. The new transformers are, in fact, not cheaper. They have certain
characteristics which we felt would work better; as it turns out, once they
were in place, we discovered that the code for reporting the port condition
was more dependent on the transformers than we had assumed.
2. That code resides in Star Command, the little DA - not in the Star
Controller. A final, retuned verson of Star Controller is being tested in
the field now.
3. We are looking for test sites on the new software. If you are
interested, give us a call. [(415) 849-2331]
Thank you,
blackey.
--
josefk@well.sf.ca.us
{apple,hoptoad,hplabs}!well!josefk
Dissemination: the upside of entropyeshop@saturn.ucsc.edu (Jim Warner) (07/18/89)
In article <12691@well.UUCP> josefk@well.UUCP (William B. Douglas) writes: > >I _am_ speaking as a representative of Farallon here,... > >2. That code resides in Star Command, the little DA - not in the Star >Controller. A final, retuned verson of Star Controller is being tested in >the field now. > If I add to my fleet of star controllers, am I going to have to remember which version of the DA to use with each of the star controllers? Are you retuning the Star Controller or the DA? I, for one, don't find your answer to be reasuring. Would you be willing to share with the net some of the details of what you're testing in the Star Command test?
josefk@well.UUCP (William B. Douglas) (08/12/89)
In article <8398@saturn.ucsc.edu> eshop@saturn.ucsc.edu (Jim Warner) writes: >If I add to my fleet of star controllers, am I going to have to remember >which version of the DA to use with each of the star controllers? >Are you retuning the Star Controller or the DA? I, for one, don't >find your answer to be reasuring. Would you be willing to share with >the net some of the details of what you're testing in the Star Command >test? Sorry for the delay on an answer, Jim. 1. No. The newer version of the DA handles both newer and older Star Controllers. 2. It's the DA that's being re-tuned. 3. We're pulsing the line with a signal that approximates the AppleTalk signal (properly, the LocalTalk signal, i.e., AppleTalk @ 230.4kbps, but we all knew what I meant, right?) and then waiting for the line to quiet down. That elapsed time is checked against a table, and the Star Command DA reports back its estimation of your port's quality. As you can see, it's targeted primarily at reflection problems. Hope that helps. If you have any questions, call us at (415) 849-2331. -- josefk@well.sf.ca.us {apple,hoptoad,hplabs}!well!josefk Dissemination: the upside of entropy