Experimental Physics and
Industrial Control System

Emmanuel Mayssat <[email protected]> · Thu, 18 Jan 2007 14:30:08 -0800

Before you send a command to your serial device, you should clear the
serial buffers.

So a read operation becomes:

1/ Read ( you will read nothing most of the time), this operation
clears /flushs the serial buffers.
2/ send your serial commnad/inquiry
3/ read the serial port a second time to get your PV value

Note that this is also assuming that you are using serial port locking,
and that this feature is working correctly (i.e. only 1 write-read
request at a time )

--
Emmanuel

On Thu, 2007-01-18 at 14:45 -0500, Phillip Sorensen wrote:
> Hi All,
> 
> I am trying to get an RS-232 ion pump controller (Terranova 751) using 
> the asyn devGpib device support.  I have the code working for each 
> command, but after time running I will get what appears to be a 
> communication timeout to the pump controller and after the timeout I 
> notice that the PV values are wrong.  They have the value of another 
> command not their own value.  For example before the timeout error the 
> voltage is 3500 and the current is 4e-6, after the timeout error the 
> voltage read 1 and the current is 3.5e-3 (3500).
> 
> The device is connected through a Comtrol Devicemaster 
> ethernet-to-serial device, and by resetting this device or restarting 
> the IOC, I can get the right values.  I have found that starting the IOC 
> inside of gdb seems to always cause a timeout with the first message and 
> trigger this problem.
> 
> By using a packet capture program, it looks like the packets are correct 
> from the Devicemaster.  The replies are to the command sent and appear 
> to be complete.
> 
> Has anyone else had issues like this?  Is there an easy way to close the 
> connection or flush the input buffer on an error?
> 
> Thanks,
> 
> Phil Sorensen
> CHESS
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System