1 May
2012
1 May
'12
1:58 p.m.
On May 1, 2012, at 12:29 PM, Chuck Guzis wrote:
On 1 May 2012 at 17:54, Holm Tiffe wrote:
The one huge advantage (depending on your point of view) of using the
FTDI devices is that the Windows drivers for them are fairly widely
installed on users' machines. On EVERY OTHER OS out there (in common
use by the Unwashed Masses, blah blah), you can just grab hold of a
USB device in userland with libusb, but on Windows, you still have to
install a "driver" for it, even if it's just a .inf file that says,
"hey, this is a userland-controlled device, just hand it over to
WinUSB" (which libusb can then use). I get pissed off when I have
to write .inf files, and users get pissed off when they have to
install drivers for a device just to use it with one application.
So yeah, I use an FT232 or equivalent a lot of time unless the raw
speed of native USB communication is required.
I don't intend to turn this into another anti-MS flamefest, so hold
off on that. This is just mind-bogglingly annoying.
- Dave
...simple or not, I for my self think that
classic interfaces are
better fit MY needs. I've finished to build an Programmer for CY7C291
and Am27C291 Chips yesterday and it has a simple AtMega644P CPU on it.
This way I have enough RAM to buffer the ROM contents + some extra for
the Program on the device. Could have build it with an FTDI in it, but
what for?
I understand. Even when working with MCUs with native USB support
built-in, I'll dedicate some pins to driving a MAX232 for RS-232
serial interface for debugging, even though it may be feeding a RS232-
to-USB adapter. It's just that 232 is simple and less prone to
coding errors. I suppose that I could go to USB using one of the
FTDI, but it's really a wash.