I think that's a different type of calibration. That calibration is done to figure out what C14 readings match to what calendar date. The calibration we're talking about, which uses controls, is calibrating the machine to ensure that what it's reading is what's actually there.
For example, water freezes at 0C and boils at 100C. But if you change altitude, the actual temperatures change, so you need to modify your scale to account for that. No controls are needed for this process. When you go to use an actual thermometer, however, you need to make sure that when it says "0C" it's actually measuring something which is 0C. So you measure the temperature of an ice cube (accounting for your altitude) and see if you get the proper result.
While i have virtually no knowledge about how a radiocarbon dating machine works, i'd like to expand a little on the issue of calibration. My background is electronics, but from what i know so far, the basic issue is the same with all kinds of machines/tools used to meassure something.
First of all the issue of a control. A control is basically just to verify that the readings you get are in the right ballpark. If i hook up my frequency counter to a crystal oscillator at 10 MHz, and my counter shows roundabout that value, i can assume that the counter basically works. It does not say anything about the absolute accuracy of either, however. It is just a way to confirm the basic correct functioning of the machine.
If one needs a more accurate test, a standard is used. In the electonics world we have resistance, capacitance and inudctance standards. Those are often accurate into the low (or even fractional) ppm range (parts-per-million). There are primary and secondary standards too, the primary one having a better accuracy.
This then goes hand in hand with calibration. Calibration is basically just a means to figure out the relationship between the expected value of an meassurement and what the machine actually reads. It does not, however, mean that the machine itself is adjusted in any way. At least not normally. You take a bunch of standards and meassure them, then record what the machine under test is reading. You then use these results to correct for any offsets/factors the actual machine has when you do "real" readings of devices under test.
In any case, to make the whole thing work you usually want a standard whose accuracy is at least one magnitude better than the accuracy of the machine you test. However, this may or may not be the case for C14 dating machines.
The next step is adjustment, if one wants it. It is often avoided, since the implications are huge. Adjustment means that you adjust the actual machine to give the reading that you expect from the standard that you meassure. The reason why this is often avoided is for stability. Any change you make to the adjustemt of, for example, a bench multimeter means to adjust physical adjustment points (like trimmers/potentiometers). But fiddling with them means that you introduce short term instabilities, since these are mechanical devices and will settle over time. The result is that right after adjustment a machine may read 100% correct, but the simple act of letting it sit for a while will cause it to drift off.
Sorry for this rambling, but i think it is really important that people understand what things like calibration or controls actually mean in the context of "machines that meassure stuff". It is not important that your thermometer reads exactly 0°C when the meassured sample is at 0°C. The important thing is that you know that a reading of 1°C, for example, is actually 0°C, and on a repeatable basis (i.e. you need to have confidence in the readings). So we (speaking of the electronics world) use controls and "low level" standards for ourselves to verify that the equipment we use is giving us correct readings (= confidence). From time to time we send the equipment to a calibration lab, which has much better standards to test the equipment against. Usually we keep records of those calibrations to see where a machine might drift to (long term stability).
While i am sure that C14 machines have quite some differencies compared to a benchtop multimeter (in means of calibration etc.), i'm also pretty sure that the general rules are the same.
Greetings,
Chris
Nice catch.
Yea, it pretty sure won't make a difference to him, especially after almost 6k pages. But then, i thought it might be a good idea to give the "average people" some perspective about the issues of what things mean when using machines to meassure stuff. Plus, i'm new to this thread (as far as posting is involved. Yes, i have read the whole darn thing!).
