I have extended the program, now a raw TCP connection to an instrument can be created.
As there is no protocol, the connection uses a waiting time of 2 seconds to receive the individual data blocks.
Link to the program is in the first post.
Peter
It might be more than 2 seconds if the command triggers a new measurement. In USBTMC, VXI11 or similar protocols you can distinguish between non-responding device and measurement delay using serial polling but in a raw TCP you cannot have this information therefore default delays should be set long.
What is the programming language you are using ? I wrote a lot of similar software in NET languages (C#, VB.NET) using various interfaces but didn't implement raw TCPIP yet.