Hi,
We were able to identify the issue and it is not in Converter itself, but rather related to the physical infrastructure. The problem is in the ARP cache on the target network gateway. After the first conversion it has cached an ARP entry for the helper IP address. When you start the second conversion, the new VM that is created has a new network card with a new MAC address, but the gateway is not aware of this change and when Converter Server tries to connect to the helper, the gateway sends the IP packets to the wrong MAC address and eventually they are dropped. When the ARP cache on the gateway expires - the problem is fixed automatically (in our environment by default it was about 15 minutes, but it depends on the specifc hardware and configuration).
A possible workaround is to generate some network traffic from within the helper VM (e. g. a ping to the default gateway IP address) that would generate some ARP traffic and refresh the gateway ARP cache. Instructions how to logon to the helper VM are available at http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&externalId=1036746. Once you login you can ping anything (even an invalid target IP address does the job) and the conversion should proceed immediately.
HTH,
Ivan