Re: [buildcheapeeg] Serial transfer data loss

From: Sar Saloth (sarsaloth_at_yahoo.com)
Date: 2002-03-14 16:14:00


><snip>
>The At90S4433 has 128 bytes SRAM. A part of it is used in the current
>firmware, but the
>rest could be used for buffering perhaps 4..5 data frames.
>Together with XON,XOFF handshake this could resolve data loss.
>
>
>Regards,
>
>Joerg

A logical assumption, but if the XON/XOFF is generated by the device
driver, then matters might get worse and not better. With Windows and
virtualized interrupts, who knows?

I was told by someone who had many computers set up (he started up an ISP
in 1995) using 486s and external modems, that Windows NT 3.51 had no
problem keeping up at 57K baud with no flow control. Things didn't improve
linearly(NT 4.0 made things much worse), but I have been told that a modern
Windows PC should not lose any characters at 115k.

I have an application with multi-channel EEG running into a serial port at
115K baud and either there is no loss or the software masks it nicely. I
am certain there is the physical channel is one-way only and there are no
hardware handshaking lines. It is running on NT4.0 at the moment. I didn't
write any of it and I don't know the protocol, so I can't say more. the
computer is a 686 at 866Mhz or something like that. I haven't heard of any
problems with it, and I think we may have a hundred or so in the field.

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com



This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:40 BST