Most parents notice, when their infant first begins to learn how to use his hands, that the child first learns to reach for whatever it is he wants with one hand. Grasping with one hand accomplished, he embarks on a period of trying to grasp everything with two hands, using the naive theory: “One hand good. Two hands better.” Eventually, the child learns that, while some things are done better with both hands, there are many times when the second hand doesn’t help, and may even get in the way.
The same phenomenon appears in laboratory automation. While more automation is always possible, there are situations where more automation is better – and there are situations where it is not.
The parallel plate rheometer (PPR) in Mark Winter’s laboratory, for example, doesn’t really lend itself to automation. A technologist has to load samples individually. It’s not that the PPR manufacturer simply hasn’t gotten around to making an autosampler for it. It is a function of the task itself.
The business end of the PPR is a pair of circular plates that rotate relative to each other, shearing a half-dollar-sized sample compressed between them. The test is to study the change in shearing properties as the material cures. The sample starts out as a putty-like consistency and cures during the test to a rubber-like material.
It is very important that there be no small air bubbles trapped between the sample and the disks. That is a visual check that needs to be done by a human.
“You could probably put enough optical sensors in to automate the PPR,” opines Winter, “but the cost benefit is not there. Not that many of these instruments (probably a few tens per year) are sold worldwide, so it is not worth devoting the resources that would be needed to automate the procedure. The instrument maker couldn’t recoup the engineering cost while keeping the selling price at a reasonable level.”
Darlene Solomon agrees that automation may not be worthwhile in terms of the complexity of the technology required in certain situations. It may not be worthwhile because the market for an automated system may not be big enough even if you made it.
There are many classes of applications (including diagnostics) where there might be a compelling reason to develop the technology no matter what the cost. For example, there is a lot of work being done in robotics laboratories to create machines to work in extremely hazardous environments. Generally, however, automation is worthwhile only when you have to do something that is general enough that a large number of people will want to buy the automated version, and it really solves problems for them.