Date: Sun 19 Oct 86 23:53:34-EDT
From: Mike Chepponis <cadre!C.CS.CMU.EDU!Michael.Chepponis@pitt.UUCP>
Subject: a couple of things.

Firstly, tnx for the experimentation - really needed!  Too bad you're not in
Pgh!

As for kiss in a single ROM, that is the the ultimate goal, but I am unwilling
to do that until I believe the code is bug-free.  I don't know why it happens,
but about 1% of the packet are corrupted with extra characters;;; I'm working
very hard trying to figure this out!  But, as I said, it should be usable as
it sits.

As for funny sounding packets, the easiest thing to do is generate a simple C
program to drive a serial port with SLIP' frames, and send out AX.25 UI frames.
That way, you can monitor them on a regular TNC.  I did this, and the audio
sounds OK to me; Be sure you send your txdelay appropriately - you might have
to make it longer for the kiss TNC (use the appropriate frame type).  And, if
you are seeing the CON LED come on, then the TNC is sending a valid frame to
the host, and if you are not seeing it, there is some problem with the host
software.  You may want to hack together a simple monitor to see what KISS is
seeing - I've done that, and I watch  145.01 here.

As a hack, I took the info you gave me about when you were having trouble
before you cut JMP6.  I think I have a ROM version that does not require you
to cut JMP6... It relies on the fact that when the address 0000 is being output
for the first opcode fetch, you said it returns an FF, which is RST 7 (call to
loc 0x38).  So, I hacked the code to do only references to 0x4000 and above
and to insert a JP 4000h at "4038h" - since an 8k ROM only goes to 3fffh, there
is address wrap-around, but by forcing A14 on, you don't have to do anything
special to the JMP6!  At least, it worked for me!  Here's v.3 of downloader
code.  Try it with JMP6 connected, let me know if it works for you, too, 
because if it does, this is the version of the downloader I should distribute!

(working my buns off, really!) -Mike