The BMI/TS4 10/27 Glitch

["Bob Scott" <rlscott@xxxxxxxx>]

I highly suspect that the core of the "10/27" problem lies with the TS4
computer not getting all the data from the datafeed (misses a bit here and
there).  One user has a W95 machine and a W311 machine and the W95 machine
gets the "10/27 glitch" but the W311 machine does not.  It turns out that
the W95 machine is very heavily used compared to his W311 machine.  This
could be the reason for the "glitch" especially if the standard serial port
which has a very small buffer is used.  There are boards made now with much
larger buffers.  Mark Brown recommends a Byte Runner Technologies board.  He
uses their model PCI -100H which has a current cost of $36.00 and which
represents excellent value and performance.  This board also looks good to
me because it is a PCI card, has a 128 byte buffer and can run as fast as
921.6 KBaud.  I suggest that this might be the best $36 you ever spent.

                 http://www.king-cart.com/cgi-bin/cart.cgi?store=byterunner

         Then select Single serial cards in the 32-bit PCI Cards column

I run my TS4 system on a dual 500 MHz PIII running Windows NT 4.0 and
essentially all it runs is TS4.  I have other computers for interfacing to
the internet etc.  Maybe this really is the common thread - running TS4 on
highly loaded computers with small serial buffers.

Regards,

~Bob