Bc2000 (for the BCF2000 & BCR2000)

--- In bc2000@yahoogroups.com, "rpcfender" <rpcfender@...> wrote:
> Strange stuff for those that might be interested, there is a reason for 
> multiple lines in the *.tx* statement.
> Perhaps you all knew this, but it was something new to me.

"Perhaps you all knew this"?
Frankly, Royce, I think that most people in this group (including me)
would NEVER assume they know anything about BCL that you don't...

But seriously: I now realize that I should probably have announced my
discoveries about multiple .tx messages more prominently than just
"hiding" them in BC MIDI Implementation.pdf vs. 1.0. But there are
always so many issues - it just got lost a bit.

> *As you loose 2 bytes for each .tx statement and I thought that
multiple 
> lines were just for neatness I made a single .tx (learn statement in 
> Mark's spec)
> *.tx  **$B0 $00 87  $B0 $20 $00  $C0 $0A*    *$F0 $41 $10 $00 $10 $11 
> $30 $25 0 0 0 0 0 val chk-1 6 $F7
> *
> but the XV3080 said that there was a bad checksum
> 
> The value after the checksum (chk-1 ) gives the starting position of
the 
> data to be summed. 6 in my case.
> I thought the BCR would look back to see the $F0 and count from there. 
> Not so, it counts from the start of the .tx message
> 
> You could count from the the start of the whole message (and make the 
> checksum offset in my case 14) or all you need to do this
> 
> *.tx  **$B0 $00 87  $B0 $20 $00  $C0 $0A
> .tx*  *$F0 $41 $10 $00 $10 $11 $30 $25 0 0 0 0 0 val chk-1 6 $F7
> *
> Now it works.
> 
> Check out Mark's spec on *Checksum Defininition*

Just one addition: as the same section in BC MIDI Implementation.pdf
explains, there is a second reason for using multiple .tx lines: any
SysEx message ($F0 ... $F7) in non-terminal position on a .tx line
leads to extra, spurious MIDI output from the BCF/BCR. (I think there
was a message strand in our group on this problem about 8 months ago
or so, but I never got round to stepping in at that time.)

So starting from your original, invalid attempt above:
1. if you put the SysEx message IN FRONT OF all the $B0 and $C0
messages, the CHECKSUM would be correct, but there would be spurious
output.
2. if you put the SysEx message somewhere IN BETWEEN the other
messages, you get BOTH spurious output AND a wrong checksum! What a
mess...

So to summarize: a SysEx message containing a "chk-" definition must
be both in INITIAL position (for a correct checksum) and in FINAL
position (to avoid spurious output). In other words: it must
constitute a single .tx line.

Mark.