NI-6602

We're using a NI-6602 board with comedi-0.7.69 and comedilib-0.7.22. We use
the ni-660x driver.  We're experiencing some problems when using both TIO
chips at the same time.

We use channels 0 until 5 as X4 encoders (channels 0-3 are located on the
first TIO chip, 4-5 on the second TIO).  When we initialize a single
channel and move that channel, everything works fine.
We've tested this with all channels.  When we use only 4 channels (only
the first 4 or the last 2), everything still works fine. When we use channels
from both chipsets, we encounter the following problem:

When we initialize channel 0 first and then channel 4, we measure on both
channels and there is a movement on channel 0, we read 0 tics on both
channels.

When we initialize channel 0 first and then channel 4, we measure on both
channels and there is a movement on channel 4, we read the expected value
from channel 4 on both channels.

When we initialize channel 4 first and then channel 0, we measure on both
channels and there is a movement on channel 0, we read the expected value
from channel 0 on both channel 0 and 4.

When we initialize channel 4 first and then channel 0, we measure on both
channels and there is a movement on channel 4, we read the expected value
from channel 4 only on channel 4. The value of channel 0 remains 0, as
expected.

The same happens for the combination of channel 1 and 5, channel 2 and 6 and
channel 3 and 7.

- We've checked the mailinglist archive and we've made the modifications
   proposed in the mail from Christian.loew_at_ni.com dated 2002-11-12.  This
   doesn't modify the above described behaviour however.
- We've also checked with another NI-6602 board and obtained exactly the
   same results.

Is there anyone who has succesfully used the NI-660x driver for
simultaneously using both chips?
Or is there anyone who has a hypothesis which could explain the above
described behaviour?

Received on 2005-03-23Z17:04:48