>>>>> "Xiaofan" == Xiaofan Chen <xiaofanc@xxxxxxxxx> writes:
Xiaofan> BTW, the current libusb version is 1.08.
Thanks for pointing me to the more recent version. When going from 1.0.2
distributed with Opensuse 11.2 to 1.0.6 from the opensuse
hardware/openSUSE_11.2 repository, this also solved the speed problem I had
with my hardware and the streaming read patch.
Setup:
Hardware delivers 16 bytes every 1.2 us to the FT2232H in synchronous mode,
as long as the FT2232H is ready to accept data indicated by TXE_N low.
Observation:
TXE_N goes high for a short time every about 40 us, probably
when 512 bytes are delivered to the FT2232H and the FT2232H switches buffer,
but after 5 buffer switches, TXE_N stays high until about 125 us from when
TXE_N when low after cthe last long period high. Probably some read request
on the USB bus got NYET and libusb decided to not query the devices any more
in that microframe.
Another hardware observation:
The FT2232H datasheet specifies 11 ns Setup to indicate write (wr_n) and
fror the data. With 60 MHz clock this means 16.66 - 11 = 5.66 ns
Clock-to-Output for the device delivering the data to the FT2232h. Spartan-6
in -2 Speedgrade even with DCM only guarantees 6.28 ns, with Spartan-3 even
without DCM reaching the constraint.
--
Uwe Bonnes bon@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------
--
libftdi - see http://www.intra2net.com/en/developer/libftdi for details.
To unsubscribe send a mail to libftdi+unsubscribe@xxxxxxxxxxxxxxxxxxxxxxx
|