As a UC Santa Barbara student, I was stationed at SLAC, the Stanford Linear Accelerator Center, working on the positron-electron project (PEP).
The PEP consisted of a circular electron-positron collider with a diameter of about half a mile. Electrons propagated in one direction and positrons, the positively charged antimatter electron equivalent, in the other direction through a beam pipe in a tunnel about twenty feet below ground. Each beam consisted of bunches of electrons or positrons that overlapped at six different points around the ring called interaction regions. The electron and positron beams crossed every 2.4 microseconds and their current was about 50mA.
When the beams crossed there was a good chance that an electron and a positron from each beam would "interact" by either annihilating into pure energy or by throwing hard photons at each other and sort of ricocheting away.
Anyway, in the experiment I worked on, a detector shaped like a cylinder lying on its side, about the size of a three-story building, surrounded the electron-positron beam so that it could track whatever happened when electrons met positrons.
When I got to SLAC, the experiment had already been running for a few years and they were just installing an upgraded version of the detector. The upgrade featured two improved components, a "vertex chamber," at the detector's center close to the point where the electrons and positrons interacted, for improved spatial resolution of particles emerging from the reaction, and, my responsibility, sodium iodide crystal end caps with fancy new electronics.
My first poignant experience in particle physics was quite similar to my first poignant experience working for a large company: Dude, these meetings are BORING.
So there I am, half-asleep, sitting along the wall of the conference room, safe from the huge elliptical adult table where professors and postdocs were trying to figure out why the new inner detector was heating up.
One guy said, "We can run the electronics for hours without any heating. But once the accelerator comes on and we have beams, it starts heating. If we let it run more than an hour it's going to melt."
Though I hadn't told anyone yet, the new end cap electronics suffered huge jolts of noise -- 2V bursts of junk every 2.4 microseconds -- the same frequency as the beam crossings. It had to be related, but the system was designed to accommodate big doses of synchrotron radiation -- hard X-rays emitted by the charges as focusing magnets direct them around the circle. What else could it be?
As for the vertex chamber, the accelerator had no thermal or electrical connection to the detector. Plus, the beampipe wasn't heating up, just the walls of the new vertex chamber. I stared at the overhead screen. It showed a diagram like this:
The detector geometry.
(Source: SLAC publication 4585; author Eliott Bloom)
Having survived graduate-level electrodynamics the year before, I couldn't look at a diagram like that without regurgitating solutions. The first thing that came to mind was the method of images. If you have a charge on one side of a conductor, the charge behaves as if there is an opposite charge on the other side of the conductor. It works for landing airplanes, too.
As bunches of electrons and positrons zoomed by, their images would have to zoom past, too. But there's a problem for the images. At the point where the inner detector closes in, the radius of the beampipe decreases by a factor of two over a few centimeters. The image current has to accelerate to keep up with the real current; a huge increase in the current flowing though the outer edges of the inner detector sees.
Without thinking, I mumbled, "The image current blows up at the discontinuity. That'll heat you up."
It generated a few frustrated looks and then they got into the schematics and started talking about a ground loop.
I'd never heard of a "ground loop" before. A more senior student sitting next to me explained what it meant and everything made sense. The "fancy new electronics" on my end caps were floating on common ground, but only connected to something resembling "earth" through the power supply. The whole end cap was nothing but a big antenna picking up RF radiation from the 50mA electron/positron current at 417kHz.
After the meeting I went down to the equipment pool and asked the grizzled old engineer who signed out equipment what I needed to simulate high frequency noise. This guy looked like a shop teacher, missing a couple of fingers, face riddled from arc welding without a mask, and he sounded like an AM disc jockey, "That ain't high frequency."
I explained what I thought was going on and he gave me a high voltage power supply. Then he took a 50Ω cable from a rack, pulled a knife from his belt, and cut it about six inches from the connector. He twisted it, taped it up, connected it to the power supply, and said, "Don't touch this part."
As I walked out of his office, more like a compound, really, he called me back. "You'll probably want this, too."
He handed me a spool of about 14 gauge wire with no insulation, just shiny steel wire.
"What's this for?"
"It's baling wire."
"You'll figure it out."
Back at the interaction region, the accelerator was down so I was allowed into the detector chamber. With a scope connected to one channel of the end cap electronics, I set the HV supply about 20 feet away, tuned it on, on and sure enough, got a big blast of noise. I switched it off and on a few more times, until I got a spark off the cannibalized cable.
The end cap electronics consisted of several hundred photomultiplier tubes housed in stainless steel. I took the baling wire, wrapped it around one of the tube housings, and then twisted the end around the signal cable's BNC connector. When I flipped the HV switch, no more noise.
I spent that night wrapping every tube in baling wire and the next day, when the accelerator came up, I got pristine signals suitable for an upgrade.
A couple of weeks later, when the accelerator was down for maintenance and people were allowed into the detector chamber, I found six professors staring at the end cap. One said, "Who wrapped wire all over the sodium-iodide array?"
I said, "There was a ground loop."
He said, "Yeah, but there's no duct tape."
Feel free to share your experiences with ground problems, lack of grounds, and ground loops.