Internet PLC Forum

I am using the Nano as a Modbus/TCP client to an Acromag analog input module.  Briefly, I open the channel, leave it open forever, and read two 16 bit registers.  I count the number of failures by evaluating the S2 and S3 status bits.  If STATUS(2) is zero, I count a no-response and if STATUS(3) is zero, I have to close the port and re-open.  If I trigger the read command every second, the number of failures over a 24 hour period is 300+ for both STATUS.  If I trigger the read command on a less frequent basis, every two seconds for instance, the numbers of failures drop to 3 or 4 each every 24 hours on both STATUS bits.
My questions: Why should I experience any failures at all?  Is there a way to issue a read command knowing that all the hardware is ready to go?
I need to query at one second intervals because I am performing Coulomb calculations. I have tested the Acromag analog module with another Modbus/TCP client and it will respond to the same query in considerably less than one second with zero failures.   As well, I cannot add any additional TCP modules until I know how to precisely time them.  I have extensively reviewed the example code supplied with i-Trilogy and there is really no substance there.

I would like to resurrect this issue.  Using the Nano with Modbus TCP servers, specifically the ADAM 6051 and a Acromag analog input module: how quickly can the TCP connection be opened and closed?  I need to sample an analog input every second and I need another two digital modules on the network as well that are not time critical.  Currently, my testing with just the Acromag indicates that this is not going to happen.  How do I manage higher speeds on local but remote IO networks?