[Gb-ccb] suggested changes to ccb library

Martin Shepherd mcs at astro.caltech.edu
Tue Aug 12 13:50:25 EDT 2003


On Mon, 11 Aug 2003, Brian Mason wrote:
>...
> The issue is that a new scan can't start until the current integration
> is done. Potentially this is a longer delay than we want.

This can be fixed at the hardware/driver level, so there is no need
for a software workaround.

Having said this, I am wondering what you consider to be too long of a
delay for a new scan to start, especially given that we were planning
on very short, 1ms integrations? Observing scans are supposed to start
on one-second boundaries, so the manager needs to inform the server
sufficiently far in advance of the target second that the driver
receives the request before that second has passed. Since the request
is going over the ethernet, and neither the manager, nor the server
are to be real-time processes, getting this reliably right will
probably require the manager to send the message at least a second in
advance.

> I believe we determined this by working through one of your (or
> Yamasaki's?) old state machine diagrams and are not sure if it's
> still true given revisions to the design.

The new FPGA design is going to be completely different, but it will
be based on the requirements that we hashed out for the previous
designs.

Martin




More information about the gb-ccb mailing list