USB, the Universal Serial Bus

 //////////////////////////////////////////////////////////////////////////////

The Universal Serial Bus was conceived as a means of simplifying connections
between computers:

 *  to allow automatic self-configuration of devices
 *  to leave behind all the worries of DTR and DSR signalling
 *  to cast null modems forever into the void()
 *  to go where no RS-232-C has gone before!

As with so many inventions, USB made a few things simple, while introducing
many complications of its own.  "Sic transit gloria teknologi."  :-)

There is a Universal Serial Bus Home Page on the web at:

    http://www.usb.org/

 //////////////////////////////////////////////////////////////////////////////

One vendor of USB-to-RS-232 converters is Keyspan:

USB High-Speed Serial Adapter
    http://www.keyspan.com/products/usb/usa19hs/homepage.spml

Mini Port Replicator
    http://www.keyspan.com/products/usb/mpr/

Solaris 10 drivers for Keyspan USB-to-serial devices
    http://www.keyspan.com/downloads-files/developer/sun/

 //////////////////////////////////////////////////////////////////////////////

Date: Tue, 18 Apr 2006 14:24:16 -0700
To: (undisclosed-recipients)
From: Ritu
Subject: Re: ?setup Hyperterminal with USB cable to connect to Filer
         console--instead of DB9 / RS232 cable?

OK: I found out the problem.

The customer was using the USB-to-RS232 cable, which has a Male DB9 for
serial connection. So, she was using a straight-through gender changer.

So, I tried various things in the lab, and the thing that works is, use
the console cable (supplied by Netapp) to change the gender.

Basically,

1. The COM ports in Device manager will show a new com port, don't look for USB.
   Upon connecting the cable into the laptop, it will want to install the newly
   found hardware. Let Windows install it automatically.

2. Plug the USB cable into the Serial Null-modem cable that comes with the
   Netapp Filer to connect to console. 

3. configure the HyperTerminal as usual (9600, 8, N 1 N)


In the lab, the above setup worked first try.

The moral of this story: the stores are full of USB-to-serial connectors
and cables. There are none with USB-to-femaleDB9--(that I could find).

(It may not have worked anyway. I could not find a straight USB-to-serial
null-modem cable either.)
 
Regards
Ritu

 //////////////////////////////////////////////////////////////////////////////

Newsgroups: comp.protocols.kermit.misc
Message-ID: <oprxw6mthj3k6cjf@news.lehigh.edu>
Organization: Lehigh University, Bethlehem, Pennsylvania
Date: Fri, 31 Oct 2003 13:21:55 -0500
From: Bryan Hodgson <RM.TAG.brhc@lehigh.edu>
Subject: Direct connection across a USB-connected DB9.

My latest laptop lacks a DB-9 serial port, useful to me for connecting to 
device consoles.  So I purchased the laptop vendor's recommendation, a 
Belkin FSU103; basically, it's a DB-9 connector at the end of a USB cable 
with some status lights to show that it's plugged in / moving data / no 
errors.  WXP.  Device Manager thinks the device really *is* a serial port, 
and not a modem.

And I dug out my copy of the Kermit-95 software, and applied the latest
(01/2003?) updates.

Microsoft's HyperTerminal works fine across this device when connected
directly to the serial port and (for example) /dev/ttya on a Sun host,
although HyperTerminal itself is still brain-dead, of course.

Enter Kermit-95.  Using a K95 console with:

Communications Parameters:
  Port: com4, speed: 9600, mode: local, modem: generic
  Parity: none, stop-bits: (default) (8N1)
  Duplex: full, flow: rts/cts, handshake: none
  Carrier-watch: off, close-on-disconnect: off
  Terminal bytesize: 8, escape character: 29 (^])
  Priority: regular

  Carrier Detect      (CD):  Off
  Dataset Ready       (DSR): On
  Clear To Send       (CTS): On
  Ring Indicator      (RI):  Off
  Data Terminal Ready (DTR): (unknown)
  Request To Send     (RTS): (unknown)

(arrived at via set line com4; set sp 96; set carr off; set flo /dir) ...

When this session is connected, the terminal displays nothing in response 
to keystrokes (e.g. enter), though though the send and receive lights on 
the serial port device both flash.  If I Alt-X back to the command window 
and C again, the terminal suddenly flushes the output of the previous 
keystrokes.

I presume that I'm clueless about some critical parameter that Kermit 
needs but doesn't have.  Suggestions would be appreciated.

 ..............................................................................

Newsgroups: comp.protocols.kermit.misc
Message-ID: <bnueu7$t1$1@newsmaster.cc.columbia.edu>
References: <oprxw6mthj3k6cjf@news.lehigh.edu>
Organization: Columbia University
Date: 31 Oct 2003 19:59:03 GMT
From: Jeffrey Altman <jaltman@columbia.edu>
Subject: Re: Direct connection across a USB-connected DB9.

The implementation of the USB Serial Port driver differs from the
standard Serial Port driver.  Kermit 95's configuration of the serial
port does not work correctly with the USB Serial Port driver.  

This is a known bug. 

 ..............................................................................

Newsgroups: comp.protocols.kermit.misc
Message-ID: <8ce22d01.0310311918.69f5448e@posting.google.com>
References: <oprxw6mthj3k6cjf@news.lehigh.edu>
    <bnueu7$t1$1@newsmaster.cc.columbia.edu> <oprxxgydea3k6cjf@news.lehigh.edu>
Date: 31 Oct 2003 19:18:39 -0800
From: Dan Skinner <JDanSkinner@JDanSkinner.com>
Subject: Re: Direct connection across a USB-connected DB9.

Jeff and Frank helped me with a problem with a win modem in a new laptop.

In the end the real solution was/is a pcmcia hardware modem.

I think this (really a pcmcia rs232 card) may be your best answer?
More expensive than a USB dongle, but cheaper than air fare to carry the
Sun machine around. Much preferable to the frustration of using Hyperterm.
Regards...Dan.

 ..............................................................................

Newsgroups: comp.protocols.kermit.misc
Message-ID: <slrnbq7ngt.7lp.fdc@sesame.cc.columbia.edu>
References: <oprxw6mthj3k6cjf@news.lehigh.edu>
    <bnueu7$t1$1@newsmaster.cc.columbia.edu> <oprxxgydea3k6cjf@news.lehigh.edu>
    <8ce22d01.0310311918.69f5448e@posting.google.com>
Organization: Columbia University
Date: 1 Nov 2003 16:23:57 GMT
From: Frank da Cruz <fdc@columbia.edu>
Subject: Re: Direct connection across a USB-connected DB9.
Bryan Hodgson <RM.TAG.brhc@lehigh.edu> wrote in
message news:<oprxxgydea3k6cjf@news.lehigh.edu>...
: 
: OK.  I tried Frank's TAPI suggestion, but that did not yield any different
: behavior.  I even found a copy of Kermit 3.15 and tried running it inside
: a DOS compatabity box (desperate measures).  That was the only potentially
: relevant suggestion I found on Belkin's web site.


Did that work?  I haven't seen the Belkin suggestion (and can't find it) but
it would be more helpful if companies who suspected problems with Kermit
software would contact us about it.

Anyway, sorry for misleading you.  I didn't realize that all USB port access
was broken; Jeff gave the right answer.  Bug #731 in:

    ftp://kermit.columbia.edu/kermit/k95/newbugs.txt

has been corrected.  As noted there, the fix must await the next K95 release.
I can't say when that will be.  Meanwhile... Thanks Dan.

- Frank

 ..............................................................................

Newsgroups: comp.protocols.kermit.misc
Message-ID: <oprx0zc3cw3k6cjf@news.lehigh.edu>
References: <oprxw6mthj3k6cjf@news.lehigh.edu>
    <bnueu7$t1$1@newsmaster.cc.columbia.edu> <oprxxgydea3k6cjf@news.lehigh.edu>
    <8ce22d01.0310311918.69f5448e@posting.google.com>
    <slrnbq7ngt.7lp.fdc@sesame.cc.columbia.edu>
Date: Sun, 02 Nov 2003 14:35:17 -0500
From: Bryan Hodgson <RM.TAG.brhc@lehigh.edu>
Subject: Re: Direct connection across a USB-connected DB9.

The Belkin suggestion for Windows 2000 was to use a DOS box.
Does not work for Windows XP.

I'm going to consider the PCMIA RS-232 option.

Thanks for your help.

Bryan

 //////////////////////////////////////////////////////////////////////////////

Newsgroups: comp.unix.solaris
NNTP-Posting-Host: 192.18.101.5
NNTP-Posting-Date: Tue, 21 Dec 2004 19:11:24 +0000 (UTC)
Message-ID: <1103656275.244009.111330@f14g2000cwb.googlegroups.com>
Date: 21 Dec 2004 11:11:15 -0800
From: frits <frits@sun.com>
Subject: USB FAQ

We have posted a FAQ on USB.
If you have corrections, updates, please let me know.

    http://www.sun.com/io_technologies/USB-Faq.html

(there is also a link on Bigadmin)

cheers
fritS

 //////////////////////////////////////////////////////////////////////////////

Newsgroups: comp.sys.sun.hardware, comp.unix.solaris
NNTP-Posting-Host: 69.158.104.50
NNTP-Posting-Date: Mon, 09 Jan 2006 21:11:31 EST
References: <dpu8jr$mj0$1@wsc10.lrz-muenchen.de>
    <UYxwf.4339$Cl2.101700@news000.worldonline.dk>
    <Pine.SOL.4.64.0601091332010.3493@marrakesh>
    <dpum7p$t2p$1@wsc10.lrz-muenchen.de>
    <T7Dwf.2042$Pq4.111399@news20.bellglobal.com>
    <vilain-2CB64F.17282209012006@comcast.dca.giganews.com>
Message-ID: <nJEwf.2936$W03.277260@news20.bellglobal.com>
Organization: Bell Sympatico
Date: Mon, 09 Jan 2006 21:11:33 -0500
From: Sunny <sunny@nospam.net>
Subject: Re: USB slooow on Blade 1000

Michael Vilain wrote:
>
> In article <T7Dwf.2042$Pq4.111399@news20.bellglobal.com>,
>  Sunny <sunny@nospam.net> wrote:
> 
>>Thomas Maier-Komor wrote:
>>
>>>Rich Teer wrote:
>>>
>>>>On Mon, 9 Jan 2006, Kent Wormsdorf wrote:
>>>>
>>>>>The Blade 1000 does not have USB 2!
>>>>>
>>>>>http://sunsolve.sun.com/handbook_pub/Systems/SunBlade1000/spec.html
>>>>>
>>>>>It only have USB 1 so the "slow" writing on a USB stick is normal.
>>>>
>>>>
>>>>Just to add to this: I've had great success with the Belkin USB 2.0
>>>>cards in my Sun Blade 1000 (at 2.0 speeds!).
>>>>HTH,
>>>>
>>>
>>>Thanks Rich--I will try to get a Belkin or at least an OHCI compliant
>>>PCI card.
>>>
>>>Tom
>>
>>There is very little to a USB2 PCI card - one chip, the port connectors, 
>>and a handful of passive components. IME there's no point in paying the 

>>Belkin premium when off brands are available at a fraction of the price--
>>ust make sure whatever you buy has the NEC chip on it and you are 
>>good to go.
>>
>>I have a 5-port NEC card branded 'A-Byte' (made in China, no 
>>manufacturer web site I could find) which cost $8.99 new and works fine 
>>in a SunBlade 100. Best price I could find for Belkin cards was $23.99 
>>(2-port) and $56.99 (5-port). AFAIK there's only one NEC USB chip, so it 
>>would appear Belkin is charging an obscene premium for the extra 3 USB 
>>port connectors!
>>
>>Sunny
> 
> 
> grepping Google's Solaris news groups for a USB2 PCI card will get you
> a big discussion on this.  The Belkin "just works".  That can't be said
> for a bunch of other "off brands".  You got one that works with your
> system.  Good for you.  At $23.99 for no hassles, know it works vs.
> $8.99 crap shoot, I'll spend the little extra.  But that's me.  YMMV.


I understand where you're coming from, and don't disagree - but the 'big 
discussion' rarely drills down to the important part - which is the 
chip. ALI and VIA chips are more common on cheap cards, but they don't 
work in SPARC systems. NEC chips do.

If you had the opportunity to closely compare a Belkin card to a no-name 
card with the NEC chip, you'd understand where I'm coming from - they 
are essentially identical, because both use the reference design 
published by NEC - right down to the PCB layout.

I'm willing to bet nobody can cite a USB2 PCI card with a NEC chip that 
doesn't work in a SPARC.

Sunny

 ..............................................................................

Newsgroups: comp.sys.sun.hardware, comp.unix.solaris
NNTP-Posting-Host: c-24-6-155-172.hsd1.ca.comcast.net
NNTP-Posting-Date: Mon, 09 Jan 2006 21:50:53 -0600
References: <dpu8jr$mj0$1@wsc10.lrz-muenchen.de>
    <UYxwf.4339$Cl2.101700@news000.worldonline.dk>
    <Pine.SOL.4.64.0601091332010.3493@marrakesh>
    <dpum7p$t2p$1@wsc10.lrz-muenchen.de>
    <T7Dwf.2042$Pq4.111399@news20.bellglobal.com>
    <vilain-2CB64F.17282209012006@comcast.dca.giganews.com>
    <nJEwf.2936$W03.277260@news20.bellglobal.com>
Message-ID: <vilain-DD72D9.19505309012006@comcast.dca.giganews.com>
Organization: DexLabs, Inc.
Date: Mon, 09 Jan 2006 19:50:53 -0800
From: Michael Vilain <vilain@spamcop.net>
Subject: Re: USB slooow on Blade 1000

No bet there.  I'm wondering how you'd buy such a beast on-line.  I 
wonder what kind of response you'd get from asking the dealer "what's 
the chipset used on the board?"  Best they probably would do is look at 
the paper documentation, if any.  It may or may not say anything.  It's 
nice to know that the NEC chips work if you run across a $9.99 special 
at the local Fry's Electronics (http://www.frys.com/ --big electronic 
retailer here in Silicon Valley known for being a geek's paradise--
silicon and corn chips, scopes and sodering irons, MOBOs,  disks,
memory, PC cases, TV, VCRs, games, and software--all under one roof).

However, at that price, it's not worth my time.  I supposed when you're 
older, you feel that way to...

-- 
DeeDee, don't press that button!  DeeDee!  NO!  Dee...

 //////////////////////////////////////////////////////////////////////////////
