[FlexNet] 6PACK TXing *garbage* in LONG bursts, ioerrs, locku ps ?

John Papson [email protected]
Tue, 3 Sep 2002 09:22:31 -0500


Maiko

re: "lh 6pack 2 /b=9600"

Assume your using standard COM-2 ?

Does port receive, if not interrupt is wrong...
Will transmit but not receive with wrong interrupt...  

Don't see entry for number of series connected tnc's. 
Does 6pack revert to 1 with no entry ? 
I'm not certain, as I always include /c=1

What high memory manager are you using ?
Have also seen some strange effects that vary upon which 
memory manager is used, particulary dates on MH list.

73

John WB2CIK

> -----Original Message-----
> From:	Maiko Langelaar [SMTP:[email protected]]
> Sent:	Monday, September 02, 2002 11:01 PM
> To:	FlexNet Mailing List
> Subject:	[FlexNet] 6PACK TXing *garbage* in LONG bursts, ioerrs,
> lockups ?
> 
> Hi,
> 
> I am experiencing something very strange. I have a very basic
> flexdigi (v3.3g) setup with one PacComm tnc (1200 baud RF, and
> 9600 baud serial port) with a 6pack eprom :
> 
> Startup is similar to the following :
> 
>  lh flexnet 200
>  lh 6pack 2 /b=9600
>  lh flexdigi
>  flex
> 
>  -----
>  
>  mycall ve4pkt 0 0
>  mo 0 1200
>  p t 30 0
>  p s 0 0
> 
> Quick question - should Flexdigi be loaded BEFORE the L1 drivers
> or not ? Some do it, some people don't. Both *seem* to work.
> 
> My digi node has reached a point where any attempt to connect
> to it (SABM) results in the digi replying with a packet burst
> in the range of 12-13 seconds in length, and not decodeable
> in my trace window (ie, garbage). I'm guessing it's trying
> to reply with a connect confirm (UA).
> 
> I am getting more ioerrs than I care to see, but the digi
> has been working fairly well despite this.
> 
> Is this a symptom of the digi simply running out of
> memory, or is this a symptom of too many ioerrs to
> deal with, and the only solution is to reboot ?
> 
> ANY help would be much appreciated.
> 
> Best regards,
> 
> Maiko Langelaar / VE4KLM
> _______________________________________________
> FlexNet mailing list
> [email protected]
> http://mailman.qth.net/mailman/listinfo/flexnet