#Telebit Modems & Amiga?
26 messages in this thread
Hello,
Actually I'm a Mac II user, posting this for a friend. He has a
Telebit Trailblazer Plus hooked up to his Amiga 2000 and although
we can connect 'FAST' mode, set at 19.2 and exchange typed messages,
We can not transfer without massive errors and aborts. I've tried
Z-Modem and X/Y Modem and Kermit. He is using Atalk and JrCom99.
Any help with AT&N modem settings and program settings would be greatly
appreciated. I suspect its a flow control problem on his end,
but we have tried Xon/Xoff and various CTS/RTS combinations to no avail…
—Thanxs (BTW: Anyone wishing to offer verbal help can call Gil at (301)
323-7695 Eves and 488-7500 (Gilbert Swann) days. He will be pleased to call
you back on the Wats… Thanxs)
Hal,
The Telebit tends to communicate in bursts which might be too fast for a
stock A2000 to pickup without errors. I'd suspect that you'd have best
results with XON/XOF protocol. I'd try locking in the modem to 9600 baud
instead of 19.2Kbaud to see if things improve. If so, the link speed is
probably the problem. I'd suggest that he try that first. My A500 is a
bit flakey on the serial end, too, but I can lock my modem in at 9600 and
track it OK.
Hope this helps. 8)
David
David,
I'll pass that on to Gil. Let you know. —Hal
It's not too fast. I've transfered files in Telebit's "fast" mode, and I'm
sending the settings to Hal via e-mail.
…Now, if I can just figure out how to use this Telebit with WHAP!….the
suggestions I've tried so far haven't worked. 🙁
Steve
Steve:
Do you know anyone locally that handles the Telebit Trailblazers? I'd
love to get ahold of one.
-Mike Schiller
In sunny Fort Lauderdale
Nope….you can always ask Brett though…. 🙂
Steve
Steve,
I suspect that "too fast" might be specific computer dependent in this
case. For example, my A500 will not reliably work at 19.2Kbaud and
regularly misses real-time MIDI data when recording with Music-X. I've
heard of A2000's that have the same problem with Music-X. Other A500's and
A2000's work fine when recording real time MIDI data with Music-X. Nobody
seems to know why. I've done extensive work on my computer to attempt to
fix the problem, but so far I'm still looking. I've never heard of people
having this type of problem on '020 or '030 equiped machines, though.
I hope that it is just settings in this case, but there is a possibility of
hardware problems. Since the symptoms were similar to my own case I though
that the possibility should be raised.
David
PS – if it is hardware related and he can get it repaired, I'd sure like to
find out about the fix! 8)
David,
I've used a Telebit at 19.2kbps several times with my A2000. I've
also had it connected to a Sun-3 via a direct serial connection at
19.2kbps. I've never had a problem with lost data. I don't know how well
it would run with MIDI, though (MIDI being almost twice as fast). What is
your Serial Buffer set to in Preferences? One hint I was given when I
started messing around with high speeds: set the Serial Buffer to the
maximum. Can't say if that would help with Music-X or not, though — I've
never used any MIDI equipment at all.
–dds
Dale,
Thanks for the suggestions. I do have my serial buffer set to the maximum
of 16k, although I suspect that most midi programs don't use the AmigaDOS
buffer. The problem seems to be more hardware dependent – works on some
computers but not others. I just used it as an example that other people
have had serial problems. In my case, all the other Amiga MIDI programs
that I've tried work fine, and my copy of Music-X works fine on other
computers I've tried it on. I'd certainly like to figure it out, but, I
find that the Amiga quite useful to me, even with my peculiar serial port
limitations. 8)
David
David,
That's weird. It's *just* Music-X that doesn't work right?
<Twilight Zone theme playing in background.> The only thing I can think of
is that there's something else doing DMA while Music-X is talking to the
serial port. Do you have a hard drive with multiple partitions on your
Ami? (If I remember right -and I may not — each partition is polled every
few seconds to see if it's been changed.) That could interrupt the DMA to
the serial port. Hmm… I suppose it could be a CIA chip that's slightly
off in its timing… Have you tried swapping CIAs to see what happens?
Mind you, this is all pure speculation, here. Still, why it should run on
one machine and not another is a real puzzle. Oh well, as long as you're
happy with it, I guess it's not much of a problem. 🙂
–dds
Dale,
No, it's not just Music-X that doesn't get along with my A500. I just use
it as an example to illustrate possible serial problems, since other people
have experienced it as well. I have tried several CIA chips & the results
are the same. It also does the same thing with the hard disk unit attached
or detached, so that rules out the HD unit. I suspect that the problem has
to do with spurious interrupts, but that too is conjecture. Thanks for
thinking about it. Somehow I think that the real solution is to get a
2500/30 or an A3000 unit. 8)
David
David,
Urm, hmm… Looks like you've had all my bright ideas first. 🙂
And now I'm fresh out. Still, as you say, it's nothing that a new A3000
wouldn't cure. 🙂
–dds
Since it's probably of no interest to anyone else, I'll e-mail you the
modem settings. Unfortunately, I don't have a manual so I won't be able to
tell you what the settings mean.
One of the registers sets the transfer protocol to use. Try looking that
one up….it might be the problem. Meanwhile, keep an eye out for the
e-mail I'm going to send.
Steve
Steve,
Could you please send me the settings as well? I can check them out at
work, where I can look at a manual.
David
Yup…sure will! Watch your mailbox. 🙂
Steve
Hal,
The 1.01 release of JR-Comm has an initialization macro for the Telebit
modem and a good amount of info on how to get the best throughput with high
speed modems on the Amiga.
-jack-
Hal,
i'e just completed a setup using a telebit trailblazer plus to connect to
uunet for a uucp connection. it took some trial and error on my part, plus
several re-reads of the manual. the amiga 2000 (mine at least) works fine
at up to 19.2 kbps through the stock serial port. tell gil to go to the
serial prefs section, and select rts/cts for the handshaking mode. this
uses the hardware connection through the serial cable to start/stop data
flow. xon/xoff just isn't reliable enough at high speed, at least that's
been my experience. the registers involved are 58=2 and 68=255. also set
a good sized buffer, say 8000 bytes. as far as the zmodem and x/y modem, i
believe that the software must be written to accomidate the "spoofing" mode
you are using. if you set register 111=20, you're telling the modem to
send x/ymodem acknowledgement messages back to the host automatically…and
just raw data to the term software you're using. if the term software is
expecting full x/ymodem blocks, and trying to calculate crc's to ack
there's bound to be a problem, since the telebit will strip the x/ymodem
info. try setting register 111=0 (no protocol), and run the transfers
under software (normal) control. then experiment from there when you get
it running. the most problems i encountered were with registers 50 and 51.
i wound up setting 50=0 (autobaud), so that it changes to whatever speed
the answering modem is at. since most modems answer at the highest speed,
then cycle down to slowest, this works well. register 51=5, which is the
highest speed my machine and term software can handle (as well as the
modem).
—Andrew
Andrew,
The problem with setting S111=0 is that the telebit does it's own error
checking, and with no protocol established, the effect is that it's making
its own blocks and adding its own crc's. So ymodem's crc ends up in a
block with the telebit calculating its own crc in addition. This
dramatically slows down throughput. Better to match protocols between the
machines and set 111 to support the selected protocol.
Pete
pete,
i understand what you're saying…have you had any problem matching
protocols with the telebit yet? i'm running uucp and the telebit g
protocol support doesn't work. the author of the uucp software
acknowledges the problem, but doesn't know where it is.
have you or anyone else solved this problem?
—Andrew
Are you using the Dillon UUCP software? When the transfer fails, what
happens? I've been having troubles trying to get my telebit to work with
the UUCP stuff, and I'd like to find out why!
Steve
steve,
well, as a test, i've created a small mail message, and sent it off with
the modem set to spoof the uucp 'g' protocol. the first (data) file gets
through with a cy (copy confirm) in return, but the software times out
trying to send the second (execute) file. the result is a failed send.
i've sent mail to matt dillon, asking if he's given any thought to where
the problem is. he had mailed me a few weeks ago that it's either the
modem firmware or uucico. i vote that it's uucico, since the modem 'g'
protocol was written for telebit by one of the original authors of
honeydanber unix. i have the afore mentioned data files as well as the
debug level output ready to send him, but i want to find out how receptive
he is first. i don't want to turn his labor of love into a misery.
the worst part of this is that the trailblazer+ loses most effectiveness if
the spoofing is not available. a tech document on the modem from telebit
says that without spoofing, the resulting piggyback of 'g' and 'pep'
protocols reduces effective throughput to just over 1200 bps, down from the
12-14000 bps that can be expected with spoofing on.
i just bought the telebit a week ago. i'm sort of bugged.
—Andrew
I've had similar problems. I've tried to send mail from my Amiga to the
UNIX system at work and although the mail arrives on the UNIX side, the
Amiga side is confused. It decides the whole transaction failed, and
doesn't delete the files in UUSPOOL:. I hope Matt figures out what is
going on.
The thing that bugs me is I'm SURE I've had this work before.
…Hmmm….just a thought. Are you using your TB PLUS to talk to
TrailBlazer Plus, or one of their newer modems?
Steve
i'm using my tb+ to talk to a t2500 on one system, and a tb+ on another.
i'm trying to get a connection through to test the system with the tb+ now
(sunday afternoon), but the system in question is down for backups.
the thing is, it shouldn't make a difference, since the spoofing is the
same.
i set the debug level to 255 and piped the output to a file…interesting
result. the first command line (telling the host i want to send the first
file) gets sent, then the first file. all is well. then the second command
line gets sent, still ok. when the second file gets sent (and the host
knows it's coming), i get no crc/ack in return. my machine attempts to
resend, and still nothing. after a few more attempts, it gives up.
all the readable parts of the output look ok, but i don't know how to read
raw control characters, so i'm stuck at that point. still no response from
matt dillon.
—Andrew
I'm using a TB+ to talk to a t2500 too, and I haven't been able to get it
to work. I think I'll try setting up another TB+ at work and see if that
works. I'll let you know what I find out.
Are you using the 1.06D version of Matt's UUCP?
Steve
Andrew,
Can't help you with that one; I work for a supplier of large multiuser unix
systems (50+ user machines) and we use the Telebit for support and
providing software updates. We use the uucp protocol without difficulty.
But I know the difference if it's not using a protocol (S111=0) or if
EITHER modem is not configured properly. Transfers then run at less than
half speed.
Pete