Hi folks I am trying to use FS as an automated test platform to send calls into an IVR system (also FS) and test the various call flows. I am using node.js to control the tests over ESL. I am using ‘originate xxx &park()’ to initiate the test call to the IVR. Then ‘detect_speech’ with pocketsphinx to listen for key phrases in the prompts from the IVR (‘followed by hash’) etc. So far so good - although I haven’t tested many different prompts yet. I am then trying to use ’send_dtmf’ to navigate the IVR menus. I am having some problems with this. After the ‘send_dtmf’ is requested, nothing happens. I read on a different post that it will work if you do a ‘playback’ before ‘send_dtmf’. I tried this, and yes, ‘send_dtmf’ now does something - but not what is expected. The CHANNEL_EXECUTE_COMPLETE event for ‘send_dtmf’ returns instantly ( e.g. 1 ms after the associated CHANNEL_EXECUTE event) rather than after the DTMF is sent, which messes up my state machine. Also the IVR system does not react to the DTMF. I made a test call to an x-lite instead of the IVR, and could hear the DTMF but it was distorted. Any help greatly appreciated. I am using 1.6.9 (installed from rpm’s) on centos 7. Thanks Phil