[Xastir] KISS setup help

[John Ronan]

On 26 Aug 2009, at 17:50, Curt, WE7U wrote:

> On Wed, 26 Aug 2009, John Ronan wrote:
>
>> Also, I'd never really seen a good explanation of the KPC3+  
>> problems before. The power cycle issue explains something I'd been  
>> struggling with trying to make a digi with a WRT54G and KPC3+
>>
>> Is it all firmware versions of the KPC3+ (Including the latest)  
>> and is the 9612+ afflicted with the same problems?
>
> Don't know on that last.  About the first, try some of these threads
> and/or searches:
>
>     08/26/2006:  [aprssig] Delayed packets (yet again)
>     10/05/2006:  [Xastir] Serial Port Questions
>     12/19/2008:  [aprssig] KPC-3+ KISS mode delayed packet issue?
>     01/25/2009:  [nwaprssig] KPC3+ in Kiss Mode
>
>     Google:  aprs "delayed packets" kiss kpc3+
>
> Thanks to kd7nm for that last bit.
Thanks,

I'm just looking at my notes here.  And this isn't going to help at  
all, but I'm going to throw it out there anyway.

I did some testing of TCP/IP and normal AX25 connection between a  
PTCIIex and a KPC3+.  Test set-up was Linux both ends.  Fully wired  
RS-232 cables.

Both radio's were on a 'quiet' channel.  I used a 8k file and ftp'ed  
forward and back it between the two machines with a window sizes of  
255  all the way through to 1785 with paclen of 255.  I ran the  
serial port speed at the highest I could set it, which was (I think  
19200) It never missed a beat.

I purchased the updated EEPROM from Kantronics and repeated the  
experiments and go the same result (interestingly though, the built  
in BBS in the KPC3+ has a hard coded window size of 2 I think.)

Some emails in the list that google turns up question handshaking,  
maybe 3 Wire RS-232 leads are in use in some places?

That said, I can stand over the above, I did it pretty methodically  
over a number of days.

The KPC3+  that we are experiencing issues with was actually on a USB  
adaptor, but the problem has now disappeared, literally, as the node  
(laptop/radio/tnc/psu) was stolen last Sunday evening.

As a complete aside, the TNC7 (Curt mentioned one recently) i tried  
to use came to a crashing halt with a maxframe of 4/window size of  
1020, it looked to me like it's buffers were filling and it was just  
sending everything in an attempt to empty the buffers. I emailed Mr  
Guttner and his reply wasn't very helpful suggesting that I was  
turning on full duplex mode in the TNC or that Linux was somehow  
doing it when initialising the TNC. Now, he may be right, and Linux  
may be at fault, but the KPC3+ and SCS-PTCIIex both performed  
perfectly under the same conditions. I've not had the time to go back  
and give it a hammering again to see can I figure out the problem.

Cheers
de John
EI7IG