Our recent equipment refreshes have all come with FDA Data Integrity, cGMP, etc. compliant by default. They have data integrity and anti-tamper features, including forced equipment self-checks (you must measure cal references periodically with internal comparison routines, if it does not pass, the system locks out), files are cryptographically signed and contain a log of any data manipulation, with the ability to always roll back to the acquired data, software versions and firmware are matched against a protocol and a certificate is generated on install, beampath accessories all have identification ICs on them and the system monitors the compatibility, again locking out if the configuration is unapproved.
These are real FDA and EU requirements, especially considering scandals like Theranos. Your 20 year old equipment will not have these features and cannot meet regulatory requirements.
And if it's human samples, now you have HIPAA and IRB cybersecurity requirements, which I guarantee you cannot meet with Mac OS 9.
Perfect example: FTIRs. The HeNe and sources all have a lifetime. Sure they're a headlight bulb and oven igniter, but if you can't get manufacturer spares and are buying the pieces off eBay, your source stability isn't traceable and therefore your measurements aren't guaranteed to not drift over time. On top of that, things like the ADC boards require factory cal.
The other very strict market for things like FTIRs is for crime labs. If the manufacturer isn't going to say the equipment is calibrated and maintained to their standards, it will rightly cause an issue in court, where people's lives are at stake.
The issue isn't whether you trust it, but whether others can trust what you did.
Here's the thing-we're not doing anything admissible to court or other regulatory agency, and frankly as an instrumental chemist I DON'T want things locked down.
Further to that, I think it's a bit rich to say that our data is "trash" because certain regulatory agencies won't accept it as correct. I've had plenty of data that I generated or helped generate published in JACS, Angewante Chemie, and subsidiary journals of their parent publishers. They don't have an issue with the data, and that's good enough for us. To be perfectly frank, sometimes we DO need to get hands-on with our instruments to do things like design experiments that you won't find published in a standard method. That's not to say that standardized methods don't have a place, but when you're trying to do things like characterize a newly-synthesized compound or observe something that was previously undocumented, often what has been written for mass consumption(such as by a regulated lab) just doesn't work for us.
To your point on FT-IR-I don't replace HeNe lasers on any set schedule, but at the same time if my results start looking iffy, it's one of those things I look at. I've put new lasers in both of my FT-IRs in the past year. The older Mattson that has a few very specific applications also got a new Glo-Bar last fall since I could not get it to calibrate correctly with the new laser. So yes, all of that stuff gets looked after.
Also, to that point, Nicolet will sell you a nice, locked down, can't tinker with it FT-IR, or a much nicer, more expensive, "research grade" instrument that WILL actually let you get down and dirty with the optical bench to your heart's content. I was trying to get the money to buy an FT-Raman earlier this year, and one of the options our Nicolet salesman(who I've known for many years, but has since retired) floated to me was a combined FT-IR/Raman that would give us a lot of flexibility. Ultimately, the money wasn't there.
Also, to that point, one of the "hot products" in mass spec right now is Thermo's Orbitrap. They've gone all in on them and have totally exited the superconducting FT-ICR-MS market, but I have serious concerns about the quality of data that those instruments turn out. I don't want to say too much publicly, but after seeing a lot of the "funny math" that goes into making the Orbitrap work, I distrust its mass accuracy beyond +/- 200 amu of ~250. At the same time, it's too much of a "black box" to really see WHAT is going on.