The environment can often largely contribute to measurement errors. Conversely, once the environment is under control, strides can be made to improve the overall measurement process. That's the basic idea behind building a laboratory for dimensional measurement: Control the environment to reduce measurement variation, and improve agreement with other laboratories.
The effort and expense of setting up a controlled environment can vary dramatically. On one extreme, you may decide it's necessary to protect a geometry machine or CMM out on the shop floor from dirt, dust, oil mist and sudden changes in temperature. In this case, enclosure will probably provide all the needed protection. However, if it is necessary to measure gage blocks or other dimensional standards, more sophisticated temperature control would be needed to obtain the desired results.
When deciding what you need, examine environmental issues that will influence your measuring process. You must consider sources of heat and cooling, such as radiators, windows, vents, operator body heat, lights, sunlight on walls and perhaps even cold north walls in the winter. You should also jot down other influences, including vibration sources and humidity, and factors such as barometric pressure if laser interferometers are to be used.
Then identify the types of measurements you will likely be doing. Are they short, simple measurements with tight tolerances—parts such as gage blocks, wires or pins? Or are they more complex—parts more than 10 inches using large measuring machines with multiple lengths and diameters, for example, that may take up to an hour to measure? These requirements provide hints as to the necessary environmental controls.
You may already have some type of controlled room in place. Is it good enough? How do you tell? First check for thermal stability. This takes some time, but it doesn't require much effort. Check for thermal drift, first over a short period—say a couple of hours; next, check overnight; and then check over a week. The basic test is the same. Set a gage with a reference dimension and check it for repeat. Do this measurement over several 15-minute increments to see if any short-term gage stabilization issues are changing the zero reading. Start the longer-term drift check. Place the reference in the gage and let it stay there. Manually or automatically record the values over time. If you don't have a digital gage, set up a video camera up to record the hand of the mechanical gage. Your movies may not win any awards for documentaries, but they should be valuable in learning what your temperature control system is doing. Also be sure to record the temperature monitor, and compare the gage reading and the temperature reading for some (often enlightening) correlations.
When it comes to providing temperature control, the information learned from your needs list comes into play. If you need to do only the basics, an on/off temperature control may be sufficient; however, mass is an excellent temperature variation compensator. A large mass provides stability despite what may be happening around it. A gage with a large steel base and a large granite or steel plate to rest the sample parts on can help compensate for air temperature changes. Perform the stability test, but monitor the air temperature and gage temperature while doing the stability check. You will likely see the air temperature vary a lot, while the gage temperature and reading may change only minimally.
If you're going to have to go to the extreme of building a top-grade temperature-controlled room, it's no place to cut corners. In fact, the cost of the equipment to control the environment will probably run about one-tenth the cost of the employees needed to operate the lab over a 10-year span. This can vary with the size of the lab and the number of people. Note that people and their associated costs are always going to be the biggest expense.
When going the route of a true controlled environment, follow the ABCs of lab layout, which include controlling the temperature around the lab; insulating the lab from the outside environment; preventing sunlight from entering; limiting access; thinking about the type of lighting used; and balancing the airflow in the room. Most controlled environments use a method that cools the air to a temperature lower than is required, then uses heaters to bring the temperature to the desired 68 degrees. This type of system is more expensive than an on-off system, but it will precisely control the temperature.
If this is the type of control you need, you probably will require help from an expert. Talk to companies with demonstrable cleanroom and temperature control expertise. They are experienced with airflow and can work with you to reduce hot or cold spots, while making the room comfortable.