If I understand, you are doing 2 voip legs. This is really not recommended because, in the least, you will create extra delay. Each time the voice packet has to be converted from voice to packet and back takes time. The more times you do this, the longer delay. Add to this any delays you have on your internet connection and it becomes worse.
As for echo, the same theory, if there is a little bit of echo somewhere to start, you may amplify it by this setup. You can check your gain settings and see if reducing the gains will help.