The source resistor(s) typically go inside the connector housing, soldered between the connector pads.
The sensitivity changes somewhat depending on the resistor value used, so using 4.7kOhm reduces sensitivity more than 2.2kOhm which reduces sensitivity more than no additional resistor. By how much I'm not certain, so in practical terms the available range may not be enough to try and match sensitivities of the mics by using no resitor on the low sens one and a higher value on the high sens one anyway. If you are on a strict budget forcing you to use what you already have and have the DIY inclination to do the resistor thing yourself, trying that may be worth a go, but your best bet will be to get two mics of the same sensitivity, no matter if you choose high or low sens.
I'd say the good options for you are these if you want to record louder stuff:
1) get two of the higher sensitivity versions, use 4.7k resistors and stick with 9V powering (batt box or preamp)
2) get two of the lower sensitivity versions and use 2.2k resistors if you want to use PIP power or 9V
I haven't bothered to install source resistors on mine, but that's only because I usually record less loud material and don't use the B3s very often. I did have some ocassional peak overloading with them on stage for a jazz trio, boundary mounted to the floor close to a drum kit and powered with Niant PFAs, but boundary mounting increases sensitivity another 6dB for direct sound and the drum transients were pretty dynamic up close so that was pushing things. If I used them more often I'd install 2.2kOhm'ers.
BTW, boundary mounting these little guys by simply taping them directly to a hard reflective surface is is a great way improve the performance of the mics for zero cost. Don't overlook this too seldom implemented technique! Stage floor, wall, monitor sides, etc can all be good surfaces.