A project manager and mentor once remarked to me that “a cable is a source of potential trouble connecting two other potential sources of trouble.” It's a lesson we all learn well, but often unconsciously suppress.
Here's my example: I recently bought a small disk drive to use as a portable backup and file-share unit for multiple PCs. It is self-powered from its USB cable, but since it draws more current than a standard USB port can supply, the vendor provides a Y-shaped cable with dual USB connectors. This lets the user connect it to two USB ports, in parallel, so it can to get enough current.
Since I found the dual-USB-port arrangement awkward and kludgey, I instead connected a separate AC/DC supply with plenty of capacity to its external DC-input jack. It was all very straightforward, I figured.
I was wrong. The various PCs didn't consistently recognize the drive, and data transfers were slow and error-prone. Since my supply could source at least twice the needed DC power, and I checked my supply's output, I was pretty sure that the supply was not the problem. Then, in a semi-mindless gesture of frustration, I connected the unused USB connector of the drive's Y-cable to the USB port of a nearby, unpowered PC. Poof! The drive worked fine, and consistently, with a variety of PCs. So I replaced the Y-cable with a standard, one-to-one USB cable, and all stayed well.
So what was the problem? As you might expect, it was the unterminated cable. One of the basic rules we learn when doing practical, hands-on engineering is that, in almost all cases, an unterminated cable is a DC, RF, EMI/RFI, noise-laden swamp. It can be the root cause of many continual problems, as well as intermittent and inconsistent ones. I knew this, forgot it, and had to remember it again. Maybe this time I won't forget it, and have to go through this cycle again, at least for a while.
-x x x-.