[aprssig] Is 9600 Baud Okay on 2Mtr's?

Henk de Groot henk.de.groot at hetnet.nl
Sun Jan 9 10:26:13 EST 2005


At 01:23 5-1-2005 -0500, A.J. Farmer wrote:
>Although I haven't tried Winlink yet, I'm not sure what all the fuss is
>about that the Kenwood radio TNC's won't do the job.  Is this coming from
>people that have tried it and failed or is this based on speculation?  I got

No, I tried it!

First of all I'm very puzzled by you message, as far as I know all the A 
models of the TH-D7 fail to work in KISS. These TH-D7's can receive but not 
transmit (on transmit you only get TXDELAY worth of flags but then no 
data). The only TH-D7 models I know that work in KISS are the TH-D7E - 
revision 2 and thet TH-D7E(G). So if you worked the AO-51 with the radio in 
KISS than you are either using an E model or maybe a recent TH-D7A(G) if 
Kenwood managed to fix it? If it is an A model then this is news - up to 
now none of the A models worked in KISS.

The TH-D7E cannot possibly work with a lot of data in 9k6 since the 
asynchonous serial link is 9600 baud while the 9k6 RF link is sychronous 
and thus faster. So you cannot deliver data fast enough. You can at most do 
one packet at a time, not good for any serious data exchange.

9k6 downlink is a desaster because of this, when a station sends you 4 
frames in one 9k6 transmission you only get the first frame, the second 
frame is lost because the radio needs the time to send the first frame over 
the KISS link. This is a showstopper for 9k6 for any connected packet use - 
unless you tell the station you connect to, to use a MAXFRAME of 1. But 
then why use 9k6 in the first place.

Sending data to the radio while receiving data from it is too much for the 
radio to bare; the TNC will hang itself up (the CON and STA icons will be 
lit, the TNC will not receive any data anymore). I got it running with a 
special version of the AX25_MAC driver that I use for DIGI_NED and it only 
works if you immediately quit transmission when you receive data from the 
radio and that reinitializes the TNC if the TNC hangs itself (there is a 
way to gain control over the TNC even it is not responding).

The radio can queue at most 2 packets, when you throw more packets at it 
than it does not drop the packet as per KISS specification, but instead the 
radio hangs itself and you have to reset the TNC.

When you send 2 I frames the radio will not send it as 1 transmission with 
2 frames but as 2 seperate transmissions with a TXDELAY inbetween. This 
happens with 1k2 and 9k6; the speed penalty is such that you are better off 
just using 1k2 (which is a lot more reliable than 9k6).

The interruption between two consecutive frames is most likely the reason 
why 2 TH-D7's in 9k6 will work together, the interruption after each frame 
gives the receiving TH-D7 the time to send the packet to the PC. Of course 
the data throughput is a joke this way.

The PERSIST value in the TNC is upside down, setting it to 255 will cease 
any transmission while setting it to 0 will always cause a transmission - 
something to be aware of when you use the radio with ohters on the same 
channel.

The TH-D7E is a PLL radio and therefore almost per definition crap for 9k6 
use. 9k6 signals contain frequency components as low as 10 Hz. A PLL will 
see components like that as frequency drift and try to compensate - this 
distorts the signal. Only a very good receiver that can compensate for this 
will be able to recieve the signal. You can tune a receiver to pick up 
these distorted signals but then that receiver will not be able to receive 
other type of radios anymore. This problem exists for almost any 9k6 ready 
PLL radio, only special data radio's do not exhibit this problem (like the 
T7F radio and TEK data TRX).

Even when not in KISS the TH-D7 behaves bad, this is because the RESP timer 
does not work. When you try to maintanin a connection with a FlexNet node 
or another node that does not put a "Final" bit on the last transmission 
(there are good reasons for not doing that), then the link will be very 
slow (the TH-D7 waits after every reception until the FRACK timer of the 
oposit station runs out and sends an RR+).

So if you think the TH-D7 makes a good radio for connection oriented packet 
use then I an very interested how you pulled this off, given the limitation 
imposed by the radio. Maybe like bob said use the TH-D7 in terminal mode to 
read bulletins on-line, but for any automated exchange, binary transfer 
etc. you better save the trouble of trying.

I hope this gave you enough techincal information what all the fuss is 
about that the Kenwood radio TNC's and why it won't do the job for any 
serious connection oriented packet use. APRS with single frames is okay, 
but don't expect much more than that.

Kind regards,

Henk.






More information about the aprssig mailing list