While performing a beam-switched observation it was noted that although both channels always had some baseline structure that in most instances channel B was not usable. The problem appears to be in the receiver room, outside the dewar, but before the fiber interface. Note that even the "good" channel appears to be less stable than one expects. This could be more than one problem but we need to understand the stability problems. The K-band receiver was removed from the telescope. After inspection a few bad connections were replaced. The receiver was then placed in the control room for testing. A plate was installed in place of the feed. One amplifier had poor 1/f noise and was replaced. Overall, the stability was improved.