The balun in front of the TI chip need/can be used e.g. to change frequency matching, which of course can improve accuracy. The circuit need then of course provide dc bias, whcih will be generated by the opamp/pwm combination, which will be controlled by the µC firmware itself. To check in detail it is necessary to compile own version of the UI (sources has been published), add a button or two to send various
CMD_SETDAC values. Then of course would be necessary to measure what CMD_SETDAC value is necessary to get 0 bias and what to get full ~23V.
I received today my TI toy, heh, and shot it immediately. Darwin awards for me, really. I don't work with MSP430, but have some launchpads here, so used one of them as SBW adapter to attached CC debugger to the SA µC. However in hurry i forgot to unselect "load program" and of course the flash got wiped :\ So i can't do any further testing, TI will for sure not provide the firmware file to me. Sure, i still can write own firmware but that was not my intention, the learn what the org. firmware is capable to do was more interessting.
Sure, TI could for sure provide some documenation or description of that CMD_SETDAC (and maybe as well the CMD_SETFREQ), but they didn't in first place so i doubt (i saw some replies from their support engineers on ti forum, ehm, yeah) they will do it now.
So yes, measure with DMM, but the best would be with as said above with an UI that have already extra button or two to send these CMD_SETDAC as well.