Rob
It has been a long time (I used to work at PacComm and did a
bit of work with the TNC-220) but if I remember right the 220
uses a 8530 as the data port device. I also believe that you
could use a "software dcd" or the squelched audio of the radio
as the dcd i.e. if there was noise on the audio line the dcd was
active. There may be just enough noise hitting the device to
cause it to hold off sending a packet, since KISS has no way
of reporting back to the host machine the status of the packets
it has received from the host, the host will load it up with more
packets and retries, then when the tnc finally does get to do
a TX it blast away with the whole load. This also may be related
to your p-persist and slotime settings i.e. how aggressive is the tnc
set to be, if it is too timid it will wait and also load up with packets
to be retried, since the host thinks that the packets never got acked,
when in truth they never got sent.
In the past some have put software detection of one of the control
lines in the host, and passed the DCD state to one of the lines on the
RS-232 cable, usually the CD line and used that signal to start and
stop timers so as to reduce this problem, I am not sure if the KISS
drivers in the AX.25 code have the ability to look at the CD line if
it reflects DCD or not. I think they TNC-220 may have a jumper in
there to route the DCD signal out to the CD line, if not you can
mod it to do so.
> -----Original Message-----
> From: Rob Compton [SMTP:[EMAIL PROTECTED]]
> Sent: Thursday, March 02, 2000 03:59
> To: [EMAIL PROTECTED]
> Subject: multiple tx frames
>
> Hello fellows.
>
> My latest problem relates to one of my TNC based ports.
>
> The TNC is a TNC220 in KISS mode, though I am sure that has absolutely
> nothing to do with the problem, but I'm sure that someone will want that
> info.
>
> The problem is thus: it transmits the same frame, probably 8-10 times, in
> one big transmission,
> and because maxframe for that port is set at 3, I quite often see 24-30
> frames of 250bytes in length being sent each time the wireless transmits.
> And at 1200bd, that's taking a lot of time on a shared channel.
>
> What could be causing it?
>
> The TNC is hooked up to /dev/ttyS1, and I don't have anything else on that
> irq (3).
>
> The TNC is connected to the port via a fully connected RS232 lead - ie: a
> 25
> core shielded cable with D-25's on both ends, passing full continuity
> checks, no shorts, etc...
>
> The other TNC based port on /dev/ttyS2 is working fine.
>
> The system is thus:
>
> Cyrix MX II PR233, 32Mb RAM. 64Mb swap memory, 4 serial ports (ttyS0-3)
> using irq's 4,3,10 & 12. USCC at 300h, PC100 at 280h, distro is Definite
> Linux v7 (a British distro based around RH6.1) with Kernel 2.2.13.
> ax25-tools-0.0.5, ax24-utils-0.0.4, libax25-0.0.7, node-0.3.0, xfbbd,
> jnos(g0mhd) 1.11d, clx501, KDE. Installed on 2 x HDD's, 1 x 520Mb, 1 x
> 1Gb,
> plus occaisional incursions from a 100Mb Zip disk.
>
> Any ideas anyone?
>
> Regards
>
> Rob G1ZPU/GB7ZPU
>
> ===================================================================
> Get paid to surf the web, tell your friends, and get paid while they surf
> too!
> Check out: http://www.teamdangerous.com
> Links, Motorsport, A look a the lighter side of Amateur Radio, Homebrewing
> and more
> ===================================================================
> << File: Rob Compton.vcf >>
