Here is something I wrote a while back. It is long, but may help.
To begin with, the "head" of a cartridge is its base or back end. That's why the markings on the back of the cartridge case are called the "headstamp".
So, headspace is simply the space for the "head" of the cartridge. In a rimmed cartridge, this is obvious, but for all cartridges, it really is a measurement of the room for a cartridge from the bolt face to whatever stops and supports it in the chamber. For rimmed cartridges, that is the front of the rim; for belted cartridges, it is the front of the belt. For cartridges like the .308, measurement is taken from a specified point on the shoulder; for a cartridge like the .45 ACP, the measurement is from a sharp shoulder which abuts the case mouth. So we say that a .308 headspaces on its shoulder, and that a .45 headspaces on its case mouth. For our purpose here, we will assume that the gun is a rifle in .308 Winchester, but we need to know that headspace is a factor in pistols and revolvers as well.
Some headspace is absolutely necessary; if no tolerance is allowed, operation of the rifle may be difficult or impossible. But while there is a correct range, headspace can be wrong in either direction. If there is insufficient headspace, a cartridge will either be difficult to chamber or will not chamber at all. In combat, this could spell disaster more certainly than excessive headspace.
What problems can result from excessive headspace? The answer is in what happens when a rifle cartridge is fired. The front of the cartridge case is made thin, because it needs to expand to seal the chamber and prevent high pressure gas from coming backward. But that thinness means that under pressure the case will grip the chamber walls very tightly. The rear of the case, being thicker, will not expand, and the pressure will push it backward as far as it can until the breechblock or bolt stops it. The case will stretch. It is nearly impossible to prevent some case stretching; if the gun is to operate normally, there must be some play between the bolt and its locking mechanism. But if the stretching is such that it exceeds the elastic limits of the case material, the case will tear apart. At best, this will leave the front part of the case in the chamber and hang up the gun. At worst, high-pressure gas will be released into the system and possibly damage the gun or injure the shooter.
Some folks confuse excess headspace with an oversize chamber, and think that excess headspace can be handled by reloading without full length resizing of cases. That is true if the case has simply expanded into an overlarge chamber, as it will do if the case head is held by the extractor or the case is rimmed. Even if headspace is excessive, and the breechblock can actually back up, neck sizing can delay the inevitable, but not prevent it. The condition will worsen with firing until no care in reloading can compensate, the case head will protrude too far from the chamber, and the case will bulge and blow out, with the pressure release wrecking the rifle and possibly injuring the shooter. No one should be deluded into the belief that excess headspace is not dangerous, or that reloading techniques will correct it.
Why are measurements needed? Why are two measurements necessary? Why not make every chamber of every gun to the exact dimensions required?
The answer involves the nature of machine work. Chambers are reamed with a tool called (surprise!) a reamer. If only one rifle were to be made, it would be possible to make a reamer to the exact dimensions and it would cut an exact chamber. But in mass production, it doesn't work that way. The designer of a cartridge specifies certain tolerances, based on his knowledge and, to some extent, the anticipated use. When a reamer is made to cut chambers for that cartridge, the reamer is made to the outside tolerance, or the largest allowable size. As chambers are cut, the reamer wears, and when it becomes dull, it is sharpened. This continues until the chamber is at the smallest allowable point, when the reamer is discarded and a new one used.
This system introduces one element of variation in cartridge chambering. The other is simple wear. When a rifle fires, the pressure generated inside the cartridge case pushes back the case, which then pushes back the bolt, which then pushes on the locking seats in the receiver. After a while, the bolt lugs and the receiver wear enough from this pressure, combined with the friction of normal operation, that the bolt can move more than desirable under pressure, and we say that headspace has become excessive.
Now, remember that reamer that was used to cut chambers? Well, it is not the only reamer involved. Reamers also cut the chambers on tools used to manufacture ammunition, and they are used and sharpened the same way, so the size of the ammunition can vary. Reloaders use sizing dies that are also made by reamers, and those reamers are made and used the same way. In factory production ammunition is made to tolerances, so some cartridges may be said to be "long" and others "short" even in the same batch.
Now, when a rifle barrel is made it is either not chambered at all, or given a "short" chamber. Unchambered barrels are used by gunsmiths to build rifles for custom cartridges. Short chambered barrels are used where the final caliber is known, but it is desirable to adjust headspace after installation of the barrel and selection of a bolt. Two gauges (or gages) are used at the factory and by gunsmiths to ensure that the chamber and bolt are within specifications for the cartridge. These are called the "GO" and "NO-GO" gauges. Their use must be understood in terms of the tolerances of the cartridges that the rifle will use.
The GO gauge ensures that the rifle will close and operate with the longest cartridge that is within tolerances for the ammunition. The NO-GO gauge ensures that the shortest cartridge that is within tolerances will not be allowed to stretch far enough to exceed the elastic limits of the case material.
But we mentioned that normal use of the rifle will cause changes in the dimensions of the locking system and the locking seat(s) in the receiver. That fact led to the development of a simple "one gauge" test to ensure that the rifle has not become dangerous. This test is by use of a FIELD gauge. A rifle that accepts a FIELD gauge may be nearing, at, or past the danger point; the only way to know which is by knowledge of that rifle, or by the "feel" of the gauge. At best, failure of the FIELD gauge test delivers a warning, like the wear ridges on tires. At worst, it signals certain danger. Even a rifle that fails the FIELD gauge test may function normally with cartridges at the long end of the cartridge tolerance, yet be dangerous with cartridges at the short end.
The term "FIELD gauge" should not be taken to mean "the field" in a military sense. No one calls "time out" in battle to check soldiers' rifles with a FIELD gauge. In this sense, FIELD simply means any place outside the factory, such as a depot or an arms room.
Another point of concern is how long a normal rifle will last, in terms of rounds fired, before headspace needs to be checked. For most shooters, the answer is, "Don't worry about it." The fact is that most rifle owners will never live long enough to see their rifles develop excess headspace. But in military service, especially in "familiarization" firing, rifles wear out rapidly, and headspace checks are routinely carried out. Match shooters too, who often fire tens of thousands of rounds a year, will check headspace every few months.
In most cases, headspace should be checked every five thousand rounds, just to be on the safe side. But the reality is that barrels will usually wear out before headspace becomes a problem, and many match rifles have had several barrel replacements with the same receiver and bolt. Since a new barrel will be final chambered on the rifle, the headspace will always be reset at the time of barrel replacement. If bolt or receiver wear makes it impossible to obtain proper headspace, the worn part is scrapped.
Jim