I figured the jump inputs on the Zephyr would no require too much current to activate them, afterall it's just going to digitize the voltage level and convert it to a DCC command to send out to the rails. So my first attempt was to simply use a 9V battery and a 10K potentiometer connected to the Jump 1 input. This worked fine, but the throttle's response was very non-linear, with all of the speed control coming in the upper ¼ turn of the potentiometer. It turns out the Zephyr is providing a 1K ohm load at the Jump inputs, so this load in parallel with the potentiometer produced the non-linear results.