Relative savings versus absolute cost
As someone who enjoys finding solutions to problems, I am frequently faced with a choice to make: pay for a turnkey solution, or cobble together my own.
Just this week, one of the computers that controls an instrument in the laboratory stopped booting. After some investigation, it was discovered that the power supply unit is broken. As is surprisingly common with expensive scientific instruments, the software is old and will not run on modern hardware. This computer also has some outdated connection hardware (serial) which isn’t standard on more modern PCs. We had previously looked into the cost of getting newer computers setup for several instruments, and the price was in the tens of thousands for the newer software, installed by a technician, onto a computer with specialised connectivity that we would have to source. It was decided to start factoring those upgrades into the next budget, and hope that the existing computers continued to compute.
Sourcing a new power supply for this particular computer was harder than I expected. The PC manufacturer does not consider the computer within warranty (which is fair, given the computer is close to 20 years old). However, they also no longer supply the specific power supply. I could, however, find some second hand power supplies on eBay. The question is now this: do I spend a day trying to diagnose and fix the power supply, or just accept the cost and order a replacement? It’s a tricky question. Power supplies commonly fail due to a faulty capacitor. If this is the case, it will be quick to find, replace (assuming I can source an exact match), and see if that solves the issue (there could be multiple failures). However, is that a worthwhile use of my time?
While spending time resolving this problem, the instrument is out of order. Projects are waiting on results, with no certainty of when they may come. In this case, it is manageable. I have other projects to work on, and if something was urgent then there is the analysis facility available at the university. However, if this was a mission critical piece of equipment, then the decision to save a bit of money eking out a few more months of use with the old computer would have been a poor one indeed.
Free software is another area which brings with it an often overlooked cost.1 If I have to spend a day digging through documentation in order to achieve a task that a paid piece of software makes easy, then there is a cost associated with that time. Apple has done very well in developing software that intuitively ‘just works’, and their market share reflects that achievement. I enjoy learning how new software works and writing custom scripts to automate data processing, but I cannot pretend that there is no cost associated with it. If I didn’t have the years of experience already, then what takes me a day of tinkering could take me weeks. Training a team of workers on a new skill (like coding) in order to use a cheaper tool quickly becomes more expensive than having paid more for the easier to use alternative.
Scientific instruments do not just need to get accurate data, they need to provide the user with an easy to use interface/workflow. Some of the cheaper brands come with confusing documentation (often in poorly translated English), and require lots of time and energy to get up and running. If there is the budget for the instrument that ‘just works’, then it is probably the better choice.
I am currently looking at SLA 3D printers. I have some ideas about microfluidic reactors, and a high-resolution printer would allow for rapid prototyping. Curiously, I am leaning toward the more ‘budget’ models. The reason is nuanced. I do not have a typical use case. I am not looking for a machine that will ‘just work’ and reliably print the files I send it. I am not wanting to rapidly prototype parts for an design team, or print highly detailed figurines. I am looking to tinker around with how the machine prints. Stretch the limits of what it can do. The high-end machines deliver a very ‘professional’ experience, with easy workflows, many safety features, and proprietary resin tanks to ensure the print will succeed. Some of the cheaper models are capable of the resolution I need, but have the added benefit of doing exactly what I tell it to do, even if that will break the machine. In this case, it is not about the budget. The more expensive machines are amazingly complex, highly engineered machines that can print fantastically well. But they can’t do what I want them to do.
Sometimes, the DIY route is the best route. If there is a specific application in mind, and no turnkey solutions available, then DIY makes sense. But DIY adds time, labour, mental effort, and troubleshooting.2 Sometimes the budget-friendly version is perfectly fit for purpose, but sometimes it will require so much maintenance that it ends up costing more than the ‘expensive’ version. The costs of downtime, maintenance, capability, and training all need to be considered before deciding on the most cost effective option. There is a delicate balance to strike, and the decision needs to be consciously made.
Sometimes it is worth paying a bit more.