From: Marko Mäkelä (marko.makela_at_hut.fi)
Date: 2005-11-30 23:01:53
Hi Levente,
> >BTW, for unlucky people like me who no longer have a serial port on their
> >main computer, there might be some hope. I've asked Patryk (silverdr)
> >if he could test his Keyspan USB serial adapter on PC hardware. That
> >adapter works fine with the C2N232 on Mac OS X. I even found a bearably
> >priced (35 EUR) local source for them. ...
>
> You might also want to take a look at this chip:
>
> http://www.ftdichip.com/Products/FT232BM.htm
You might also want to take a look at this page:
http://www.iki.fi/~msmakela/electronics/relay/#serial.gpio
It documents my frustration with the FTDI drivers. The unsupported Linux
driver and the supported Win32 driver have different sets of bugs. There
doesn't seem to be an active maintainer for the ftdi_sio driver in the
Linux kernel, and the FTDI support only provided RTFM type answers without
actually reading my questions, assuming I have done something wrong.
> I'm working on an interface design (in principle, kind of similar to
> your c2n232) incorporating this device. This is one of the chips
> commonly used in those USB-RS232 converter cables.
I hope you can get XON/XOFF handshaking to work. It doesn't appear to
be at all supported in the Linux ftdi_sio driver. The Win32 driver does
that, but it has some bugs that prevent c2nload and other programs from
working. For instance, if you set a write timeout, WriteFile will falsely
report that it wrote 0 bytes.
Also, you can't do bit-banging on the serial lines, because the driver will
incorrectly cache the results. Bit-banging works on the Linux ftdi_sio
driver, but very slowly, because every operation results in the exchange of
at least two messages on the high-latency USB bus. There's a separate
API for general-purpose bitbanging, but I didn't want to waste time in trying
it out, as it would be specific to the FTDI chips.
I have also had bad luck with a PL2303 based adapter, on both Mac OS X
and GNU/Linux. I didn't test it on Windows.
Marko
Message was sent through the cbm-hackers mailing list
Archive generated by hypermail pre-2.1.8.