[comp.dcom.modems] 'Blazers and slow response

pjh@mccc.UUCP (Peter J. Holsberg) (02/06/88)

I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
I can type faster than the 'blazers echo the characters to my screen.

I use Pro-YAM on the clone with the Blazer at default settings.  The
Blazer on the 3b2 is set up as per MIchael Ballard's recent postings.
Anyone have any idea what's happening and how to correct it?

Other symptoms: the screen scroll is quite jerky; ^C is (frequently)
ignored; ^S stop sthe scroll but ^Q will not restart it, nor does
sending a BREAK have any effect.

Thanks.

-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800

pavlov@hscfvax.harvard.edu (G.Pavlov) (02/06/88)

In article <188@mccc.UUCP>, pjh@mccc.UUCP (Peter J. Holsberg) writes:
> I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
> Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
> I can type faster than the 'blazers echo the characters to my screen.
> 
  As has been pointed out before, the Trailblazer is not a true "full-duplex"
  modem....

   greg pavlov, fstrf,  amherst, ny

pjh@mccc.UUCP (Peter J. Holsberg) (02/07/88)

In article <510@hscfvax.harvard.edu> pavlov@hscfvax.harvard.edu (G.Pavlov) writes:
|In article <188@mccc.UUCP>, pjh@mccc.UUCP (Peter J. Holsberg) writes:
|> I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
|> Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
|> I can type faster than the 'blazers echo the characters to my screen.
|> 
|  As has been pointed out before, the Trailblazer is not a true "full-duplex"
|  modem....
|
|   greg pavlov, fstrf,  amherst, ny

Sorry, but I must have missed that.  What is the blazer doing if not true
full duplex?

-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800

grr@cbmvax.UUCP (George Robbins) (02/07/88)

In article <188@mccc.UUCP> pjh@mccc.UUCP (Peter J. Holsberg) writes:
> I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
> Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
> I can type faster than the 'blazers echo the characters to my screen.

The Trailblazers "packetize" that data resulting in some echo delay.
This is something one can usually adapt to in exchange for the overall
improved thruput.

> I use Pro-YAM on the clone with the Blazer at default settings.  The
> Blazer on the 3b2 is set up as per MIchael Ballard's recent postings.
> Anyone have any idea what's happening and how to correct it?
 
> Other symptoms: the screen scroll is quite jerky; ^C is (frequently)
> ignored; ^S stop sthe scroll but ^Q will not restart it, nor does
> sending a BREAK have any effect.

Using flow control and a higher interface rate, may result in some
jerkyness and delay in responding to control/c and other controls that
require action from the remote system, since the modem is buffering
data and doesn't know when to "dump" unwanted output data.  In "text"
mode, you might want to try using S68=4 to enable both hardware and
control/S - control/Q.  Also make sure that you're control/Q isn't
being eaten by the BIOS or communications programs.

What are your flow control settings on the modem on each end?  It is
possible that telebit is doing something nasty...  Having the modem
"pass" control/S's after the first one to the host system, but "eating"
all control/Q's in the modem would be fatal.

TELEBIT:

	You should add a mode S56=0 which means that *any* character
received from the host reenables transmission.  This mode is used by
many interactive systems in place of the strict control/Q - control/S
discipline.

In general:

Remember folks, control/S - control/Q isn't a line discipline, it is
a control sequence for running paper-tape readers on model 33 teletypes.
There are a lot of potential problems where more than one agent in a
communications linkage is trying to interpret and correctly process
control/S - control/Q.

-- 
George Robbins - now working for,	uucp: {uunet|ihnp4|rutgers}!cbmvax!grr
but no way officially representing	arpa: cbmvax!grr@uunet.uu.net
Commodore, Engineering Department	fone: 215-431-9255 (only by moonlite)

dbuckler@hpihoah.HP.COM (Dan Buckler) (02/08/88)

/ hpihoah:comp.dcom.modems / pjh@mccc.UUCP (Peter J. Holsberg) / 11:55 am  Feb  5, 1988 /
I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
I can type faster than the 'blazers echo the characters to my screen.

I use Pro-YAM on the clone with the Blazer at default settings.  The
Blazer on the 3b2 is set up as per MIchael Ballard's recent postings.
Anyone have any idea what's happening and how to correct it?

Other symptoms: the screen scroll is quite jerky; ^C is (frequently)
ignored; ^S stop sthe scroll but ^Q will not restart it, nor does
sending a BREAK have any effect.

Thanks.

-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800
----------

pjh@mccc.UUCP (Peter J. Holsberg) (02/09/88)

(Dan,

Were you going to respond to my posting?  All that happened was a repost
of the original article.)

-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800

karl@ddsw1.UUCP (Karl Denninger) (02/10/88)

In article <510@hscfvax.harvard.edu> pavlov@hscfvax.harvard.edu (G.Pavlov) writes:
>In article <188@mccc.UUCP>, pjh@mccc.UUCP (Peter J. Holsberg) writes:
>> I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
>> Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
>> I can type faster than the 'blazers echo the characters to my screen.

>  As has been pointed out before, the Trailblazer is not a true "full-duplex"
>  modem....

How old is your Trailblazer?  More specifically, what revision firmware is
in it (at *both* ends)?

We received two of the new ones recently (BA4.00 proms) and there is nearly 
*no* turnaround delay I can see.   I use these things interactive all the 
time -- beats the heck out of a 2400 baud connection! 

On hitting a key via the TB, there is a barely noticable pause before the
character echos back.  Note that I'm going out via 'cu' and 'xcomm', which
incur delays of their own in processing the characters -- my comparison is
with a local line on a real terminal *directly* into the CPU.  Our units
exhibit no more delay than you would expect if the data were flowing at 2400
baud with regards to interactive echoback (realistically, I doubt it's even
100 ms... it really is *barely* noticable).  By the time your finger leaves
the key the character has echoed.

Of course, when you get a nice full page of output, it's so much faster that
there's no reasonable comparison to be made.

We do talk to an older model for our newsfeed (Rev BA3.??), and it has a 
more noticable turn-around time.  Our units also seem to have a problem 
once in a while with that rev of the TB+ -- occasionally it will retrain 
several times during a call, then give up and disconnect.  This is not
consistent, and may be just a broken modem on the other end (it doesn't
occur with any of the other TB+ modems we talk to, and we talk with
several).

If you're not current on the firmware, have Telebit update the thing (I
would presume this is just a PROM change, right?)

------
Karl Denninger		       |  Data: +1 312 566-8912
Macro Computer Solutions, Inc. | Voice: +1 312 566-8910
...ihnp4!ddsw1!karl	       | "Quality solutions for work or play"

clif@chinet.UUCP (Clif Flynt) (02/11/88)

In article <188@mccc.UUCP>, pjh@mccc.UUCP (Peter J. Holsberg) writes:
> I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
> Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
> I can type faster than the 'blazers echo the characters to my screen.
> 


The folks at Telebit can answer this better than I, but...

  I just talked with them at their Uniforum booth yesterday, and I was
told that the new ROM for the Trailblazer can use a smaller packet, and
that the typeahead problem pretty much goes away with that.

  The problem seems to have been caused by the modem waiting until it
got 'n' characters before transmitting.  Now 'n' is smaller, and the
delay is less noticeable.

  I played a bit with their demo unit, and couldn't get more than a 
character or two ahead of the echo, and I notice that much typeahead
on my 2400 B modems' when the host I talk to is busy.


-- 
------------------------------------------------------------------------
My Opinions are my own. I can't imagine why anyone else would want them.
Clif Flynt	ihnp4!chinet!clif
------------------------------------------------------------------------

pjh@mccc.UUCP (Peter J. Holsberg) (02/11/88)

AHA!  Ours have the 3.?? ROMs.  I believe that a call to Telebit is in
order!!  (Incidentally, we ordered and received ours at the end of
December from a distributor.  It's possible that we got his old stock
and not newer stuff direct from TB.)

Thanks for the note.

-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800

pjh@mccc.UUCP (Peter J. Holsberg) (02/14/88)

In article <2338@chinet.UUCP> clif@chinet.UUCP (Clif Flynt) writes:
|In article <188@mccc.UUCP>, pjh@mccc.UUCP (Peter J. Holsberg) writes:
|> I have a Telebit TRailblazer on my 3B2/400 and one on my XT clone.
|> Whenever, I call the 3b2 from the XT, I experience a strange phenomenum:
|> I can type faster than the 'blazers echo the characters to my screen.
|> 
|
|The folks at Telebit can answer this better than I, but...
|
|  I just talked with them at their Uniforum booth yesterday, and I was
|told that the new ROM for the Trailblazer can use a smaller packet, and
|that the typeahead problem pretty much goes away with that.
|
|  The problem seems to have been caused by the modem waiting until it
|got 'n' characters before transmitting.  Now 'n' is smaller, and the
|delay is less noticeable.
|
I spoke with them yesterday and got a similar answer.  However, unless
you took delivery in 1988, the new ROMs are $99 each!!

-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800

jordan@zooks.ads.com (Jordan Hayes) (02/17/88)

Clif Flynt <clif@chinet.UUCP> writes:

	The problem seems to have been caused by the modem waiting
	until it got 'n' characters before transmitting.  Now 'n' is
	smaller, and the delay is less noticeable.

Nope, PEP is time-based, not character based ... the modem waits "n"
msec. before the line gets turned around ... now "n" is smaller.

/jordan

pjh@mccc.UUCP (Peter J. Holsberg) (02/17/88)

In article <2419@zodiac.UUCP> jordan@ads.com (Jordan Hayes) writes:
|Clif Flynt <clif@chinet.UUCP> writes:
|
|	The problem seems to have been caused by the modem waiting
|	until it got 'n' characters before transmitting.  Now 'n' is
|	smaller, and the delay is less noticeable.
|
|Nope, PEP is time-based, not character based ... the modem waits "n"
|msec. before the line gets turned around ... now "n" is smaller.
|
|/jordan

Well, that's NOT the problem I'm having.  Fairly often, the system seems
to just stop and doesn't send or accept characters -- for seconds at a
time!  In any case, 300 bps is 30 char/sec or about 6 words/sec = 360 
words/minute.  And I sure don't type that fast!  I see MUCH more delay
than that.
-- 
Peter Holsberg                  UUCP: {rutgers!}princeton!mccc!pjh
Technology Division             CompuServe: 70240,334
Mercer College                  GEnie: PJHOLSBERG
Trenton, NJ 08690               Voice: 1-609-586-4800

larry@tapa.UUCP (Larry Pajakowski) (02/21/88)

I've noticed the long pauses also but only on verr bad lines.  I suspect that
the error correction is attempting multiple times.  Using the "least cost
routing" (read cheap lines) our PBX provides results in some very bad lines to
a small town in central Oregon.  Sometimes the Blaizer won't even make the
connection (are the lines to Chile better?).  This is the only times I've
noticed the slow response problem.  BTW I usually abort the call and replace
it with usually improved results.