[quote="pbreed"]
>Submit a support request with the configuration info.
I'm unsure how to initiate a "support request".
>How the board is configured.
The overall system consists of 4 radio modems with serial output driving 4 netburners cabled via a network switch to a linux box running 4 instances of Intercepttty to create virtual serial ports.
This is because the application used to run on hardware with multiple real serial ports.
The application runs successfully using a multiport comtrol devicemaster pro device server. But we want to migrate to netburners for size, temperature spec and cost reasons.
The radios we're connecting to have a main data port 115k,8,1,n RTS/CTS.
Also a setup port tx,rx only at 19,200. We're wired to Netburner port 0 for the main data port and port 1 for the setup port. Serial settings match device settings as above. We can telnet into the radio's setup port via netburner port #1 with the common IP port 23 no problem.
The issues are with port 0 of the netburner interface with the radio's main data port.
Fixed IP address 10.51.10.93 (for example) on netburner
Port 0 settings:
Never make outgoing
9036 target port
10.51.10.23 target machine
UDP
default packetization.
>The software version number.
Most of our testing used 1.54 the factory load.
We have tried 1.59 and believe it doesn't work at all.
We just tried 1.58 and it seems to work with outward symptoms same as 1.54
(see below)
>The IP addresses of the device and the server its trying to attach to.
See above. The software consuming the serial data runs on a linux box with IP address 10.51.10.23 and looks
for data on port 9006, 9016, 9026 and 9036 from the 4 netburners on the test network.
>Are you using the device DNS name or the IP Address to connect to.
Fixed IP addresses all around.
>As much info as you have.
If we power up at a time when no serial data is being presented to any Netburner,
the network seems to come up as we had hoped. Then when the 4 radios start hearing messages they seem to get properly packetized by the netburners and arrive at the destination ports OK.
We are running multiple instances of intercepttty on a linux box to process the data with a
software application that was designed to work with multiple virtual serial ports.
However, if someone power cycles one of the units [radio/netburner pair] while over the air transmissions are occuring (i.e. serial data being presented to the netburner while it is booting up) then something hangs and no data from that unit will ever show up back at the linux box.
We have tried all possible combinations of handshaking on the netburner and on the radio with the same result. We have, however, discovered that if we go in to the web setup page "serial page" handshaking setting and change the CTS to NONE on netburner port 0 hit accept and then change it back to CTS and hit accept a second time, then data will flow. Only this particular sequence of "double whammy" will get the serial data flowing. It takes changing the setting and then changing it back to snap the interface out of its hang state.
Needless to say, we can not have a human operator having to execute this double setup operation in the deployed system as it will have at least 10 radios streaming data by way of 10 netburners that will be miles away. We cannot require a fixed power up sequence either because the units will be in different locations and subject to different electrical power interruptions.
We have tried 1.54, 1.58 (appear the same) . Tried and 1.59 can not be made to pass any data
no matter what kind of "accept" whammy tricks we try.
>If your DHCP is slow its possible that the device is trying to send too soon, thus waiting for the retry time interval before attempting to connect again.
>If you let it sit for a bit does it eventually conect?
We are using UDP and not TCP. We are not using DHCP so the above questions do not appear to apply.
We would like someone to take a look at the firmware. We believe the serial port needs some kind of a reset AFTER the netburner completes its bootup sequence or the port needs to be blocked during the bootup sequence that is not happening currently.
Any help or ideas will be greatly appreciated.