Process monitoring takes more and more time; and it costs more and more money. Examples include air in the room and air that is released to the environment, water before during and at the end of the process, assorted liquid chemicals for cleaning, metalworking fluids, other process chemicals, process gasses, and production equipment. Collecting and interpreting the data can be daunting. What does it all mean? Even if you do 100% inspection, I’d like to suggest adding simple monitoring with witness samples.
I’m using the term witness samples in a broad sense to include any kind of coupon or object or device that emulates the product.
You can use witness samples in cleanrooms and controlled environments as well as in a general manufacturing facility.
Many of us monitor air quality, particularly in cleanrooms or controlled environments. In an overwhelming number of facilities, monitoring air quality primarily means checking the particulate level. It’s essential to monitor particulates; sometimes particulates in the air are specified in the contract. But, what actually reaches cleanroom surfaces, including benchtops and product surfaces? Most facilities do not monitor thin film contamination in the air. However, even if you are concerned primarily with particulate levels on surfaces, thin film contamination contributes to particle adhesion.
Witness samples are helpful in figuring out what’s going on in the environment overall. One simple witness sample is a glass plate. It’s relatively simple to place such plates at various positions within the cleanroom. Witness samples can be an easy way to check not only for general industrial particles and dust but also for process fluids, greases, and sometimes even fuzzy clothing, sneezes, and bits of snack that shouldn’t have been in the manufacturing venue in the first place.
Particularly if the substrates (the materials being cleaned or assembled) are very sensitive, it can be instructive to use those same materials for the witness sample. This allows you to detect subtle materials compatibility problems.
Is a witness sample too sensitive?
A witness sample is supposed to be sensitive; it is often designed to exaggerate the problem. You might choose to expose a witness sample to the environment for days whereas the product itself might be carefully stored, except during assembly. Witness samples give you an exaggerated, concentrated idea of contaminants. It’s easier to see potential problems. You want the witness sample to fail before the product fails. Seeing can mean a visual observation, with or without magnification. Seeing can also involve analytical observations.
What about using witness samples outside of a cleanroom, on a factory floor? Well, it depends on the situation. If you’re not in a cleanroom, the product can probably tolerate a bit more exposure to contamination. However, even in general manufacturing, avoiding excess contamination is a good idea. Sometimes, we see processes set up near an exterior door or a loading dock. If there is a yield problem, having a few witness samples around can help you track potential problems. Yes that dust storm might just have contributed to product yield issues! Perhaps you need to shut the door during the company barbeque! Witness samples may provide an indication that investment in a storage cabinet would be a wise move. Performing cleaning immediately after fabrication and then storing the product can be a cost-effective approach to manufacturing.
You can also use witness samples to monitor manufacturing processes. In such instances, it is good to use similar materials to the product. Especially in cleaning processes, the witness sample ought to emulate the shape of the product, but it should cost less than the product. It could be the product itself, such as a unit that fails inspection due to a dimension that is out of spec (i.e. a reason other than cleanliness) and would otherwise be scrapped. Even if you have 100% inspection, witness samples can be used for destructive testing.
Not a magic solution
Using witness samples won’t solve all of your process problem; and it is not meant to replace other monitoring programs. However, witness samples can provide a bottom-line summation of the impact of the process on the product. Witness samples can also help you track the environment. If you have multistep processes, if there are so many variables that you cannot possibly monitor them all, consider where using witness samples can help you cut to the chase.