Das16 driver with PC104-DAS16JR/16 and all-Intel PC104 single board computer

Hello all,

 I posted to the Comedi newsgroup last September with some problems getting
the comedi drivers to work on a Geode CPU.  I am very grateful for the help
I received then.

 The solution to the problem, for me at least, was to return the Geode based
computers, and go with an all Intel system.  Even now, however, I cannot use
the newer versions of comedi (I have comedi-0.7.65 working).  Perhaps there
is a link with the problems I'm seeing now.

 When acquiring (8 channels, at 16kHz), every 30 or so minutes (not the same
every time) I get a "glitch" on all 8 channels.  The glitch lasts about 3-4
samples on each channel, and then everything returns to normal.  It doesn't
look like lost buffers, because my channles are becoming out of sync, and
the amount of data (3-4 channels) is much smaller than a single DMA read on
the card.

 Here are some other problems, which may or may not be related:  If I use
the newer version of comedi, I get frequenc missing buffers (the channels
all become out of sync, every 4 seconds or so).  I cannot use DMA interrupt
mode, even on this Intel motherboard (I use the timer mode written for the
Geode boards, in the Das16 driver).

 I've traced through the code down into the das16.c driver, and I can't find
any reason why there would be glitches, or why the newer comedi from CVS
does not work.

 Any ideas or suggestions would be much appreciated.

The system I'm using is an Advanced Digital Logic MSMP5SEV pentium PC104
single board computer
(http://www.adlogic-pc104.com/products/cpu/pc104/datasheets/msmp5sen-sev.pdf
) with a ComputerBoards PC104-DAS16JR/16
(http://www.measurementcomputing.com/cbicatalog/cbiproduct.asp?dept%5Fid=137
&pf%5Fid=522&mscssid=FFF82VMXV57F8HX7PW63RW1CKP4FENRE) (with a 512 byte
FIFO).

Thanks,

-Peter

--
Peter J. Jensen