Control and status information between a mobile device and another device can be signaled over a voice channel using DTMF tones, e.g., a pre-defined sequence of DTMF tones can signal a desired message (more generally, a feature code). Tones sent are separated from each other by a time gap (Tgap timer). If a receiving device does not receive a sequence of DTMF tones that can be matched to stored definitions of feature codes, then the receiving device generates a negative ACK. Also, if the sending device does not receive either negative or positive ACK, then sending device resends all tones of the feature code after a retry delay (Tretry timer). Rather than having the Tretry timer start when tones for a feature code start to be transmitted, the Tretry timer is started after the Tgap timer expires for the last tone of a feature code. Thus, only one timer is running (either Tgap or Tretry), rather than having multiple timers running concurrently. Race conditions between the timers are avoided.