In a properly wired DI box, the shield of the unbalanced connection is wired to pin 1 of the XLR. Since the shield is signal-bearing, this would induce a voltage over pin 1 of the XLR line. A ground lift switch breaks this connection, but why would you want it to begin with? Wouldn't you always want your ground to be at 0V, and not carry a voltage during normal operation?
Tyler you don't quite understand what's inside a direct box. The shield of the unbalanced input does not have to be passed to pin one of the XLR. The tip and sleeve of the unbalanced input feeds the two inputs of the primary of the DI transformer. While the sleeve is also tied to pin one through the ground lift switch. When you get a ground loop with a lot of hum & Buzz, you hit the ground lift switch to lift the ground.
Even an unbalanced output can be treated and fed as a balanced output. That's because an unbalanced circuit has an alternating current/voltage between hot and ground. When that's fed into a transformer primary, and the voltage is then induced into the secondary of the transformer. And really that's all that's required. There is no physical connection between input and output on the transformer. Transformers provide for electrical isolation between devices. Now if that ground sleeve of the unbalanced input wasn't connected to the primary of the second connection of the transformer, you'd have nothing. But that's not how it's wired.
Transformer less active direct boxes handle this a little differently. Electronic components are utilized to simulate the differential input of a transformer while converting the impedance through an FET. Since it requires phantom power to operate, pin one of the XLR cannot be lifted. The ground lift on an active DI is handled differently.
What's difficult to understand is that the ground connection still exists on both sides. They are just not tied together within the DI box. So ground is still referenced to ground even if it is floating. Both pieces of equipment are grounded so they don't need to have an extra ground wire running between them through a DI box. Sometimes it's necessary when a piece of equipment has no reference to ground to keep the ground lift switch off.
When I've wired control rooms up, the rule of thumb has always been to ground at the source. The only time it's really necessary to have a ground wire connected to ground on a microphone cable is to be able to pass phantom power. Otherwise in many balanced circuits, pin one is grounded at the source and is not actually connected at the input of where it's going. It's not necessary in that situation because there is no audio on the ground. It's a differential balanced circuit. That only requires two wires both of which are hot and are 180° out of phase to each other.
So this gets confusing when trying to deal with DC theory and AC theory. Both of which are factor in the land of audio. Most amplifier circuits naturally pass DC. You don't want that in any of your other equipment. So the DC is filtered out via capacitors to allow only pure AC to pass out or in. Analog tape recorders could never record DC. Digital can. But we don't want that in our audio. It's hard to make a speaker that can only travel in one direction and still be heard.
I work with both active and passive DI boxes and active and passive splitters. Believe me, I much prefer passive transformer DI boxes most of the time and passive transformer splitters all of the time. I hate those active splitters. And I only use an active DI box when I need to plug it into passive guitar pickups directly. The active DI boxes utilizing FET's have an extremely high input impedance like that of a tube amplifier. The transformer DI boxes have a maximum of 50,000 ohm loads which is a bit low for passive guitar pickups. It's also the only thing we had to use 30 years ago and it worked out just fine. Though there is an audible difference. I always take bass guitar direct. When I take a standard six string acoustic or electric guitar direct, I generally want an active DI box and a 57 to put on the amplifier speaker also going to 2 separate tracks.
While you might be confused by the grounding issue, there is no doubt that these DI boxes both passive and active work. Whether you are confused by the potential of the ground or not. And the ground lift switch is an absolute important necessity that works. Ground on an unbalanced feed has to rely on the active hot side to ground. But ground is still at 0 V potential even though it's part of the alternating current. By then it's different for a balanced circuit. A balanced circuit does not require ground. Only for Phantom power. The microphone inputs on my patchbay to my Neve, while they have phantom power going to the microphones, the 1/4 inch input ground sleeve in the patchbay, is not connected to ground. The ground is passed through on the back of the patchbay to the microphone connector to reference 0 V DC for the DC Phantom power to the microphone. On the patchbay, you can not plug-in a condenser microphone because there is no ground on the 1/4 inch sleeve. You can however plug-in a dynamic microphone which will work just fine without any ground connection to the console preamp, at all.
Many companies today are making all sorts of newfangled direct boxes. And many different types of electrical concepts are utilized. All of which still have to be able to lift ground. They all have to comply with UL, it's that simple.
What? You don't believe in voodoo and black magic?
Mx. Remy Ann David
So what really is your question and why? How is this affecting your recording process?