366
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 05 Feb 2024
366 points (89.6% liked)
Technology
57226 readers
4324 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
On the nose.
I used to love the bleeding edge, then my father (retired engineer) enlightened me on why important (electro-mechanical stuff) runs on older, slower, (but insanely reliable) engineering.
It's that insanely reliable part. Kind of a hare vs the tortoise kind of thing. It's more important to be able to predict when the tortoise arrives, than to be unpredictable like the hare, even if the hare finishes first 90% of the time. That last ten percent could be a massive cost.
Look at the ECU in a car - over the 40 years I've been working on cars (and my brothers and friends), we've seen exactly ONE ECU failure, and we think that was caused by an external event (a voltage spike).
I've bought a few "new tech" solutions only to have the company disappear within a couple years. For example, software for replicating a Windows install that could then install on any computer, retaining all the config and software. It was intelligent enough to update drivers as needed. They were around for 2 years, and the license has to validate against their servers. Bastards. Lol. (I'm guessing Microsoft acquired them to keep people from using it).
And that's also why it was designed to need their servers to authenticate against: because they could charge Microsoft more if their product could be switched off remotely. They likely built the product with the aim of getting bought up. Who wants to run a company for 40 years when you can just skim a few million off of Microsoft and retire?
Lol, probably true.
They know how valuable the tech was. I still have the software and all the licensing info. Part of me feels like doing a Wireshark to see what it's trying to do. I should've done that while they were around, to see the traffic.
I am in the controls and automation sector. Your dad is half right.
You have to look at the whole system to determine reliability. Oh sure what you have now is going to last a long time but no one knows how to repair it when it dies. Plus all the components are no longer available so now a repair job is going to involve lots of work. I can't count the number of times I have had to bail out a plant because the one guy who knew how it worked retired. Very expensive.
Why do you need reliability? No really ask yourself this. Or better yet how much do you need. The answer is not going to be "as much as possible". Are you going to do this process this way for 30 years? No? Ok why did you pay 4x as much for a machine that can last 50? There is almost no chance whatever process you are doing is going to be the same forever. There will be improvements and your competition will take advantage of them.
What the heck does reliability even mean? Take a classic FET output vs a relay output.
Relays: easy to repair, can handle almost any voltage you throw at them, sensitive to vibrations causing false state changes, rated for a 10k cycles.
FETs: can't repair, can handle limited voltage ranges, ignore vibrations, rated for a million cycles.
Now tell me, given these facts, which one will result in less downtime over say ten years? I am asking here because I don't have enough information. Again you need to look at the whole system to determine which way you want to go. I have seen machines act up seemly randomly because a relay was flipping from vibrations and I have seen ones fuse contacts causing a machine to rip itself apart. I have also seen FETs die from a tiny surge.
Maybe you don't want a repairable system. I know this might be startlingly fact but it might benefit you to not have any ability to mess with your machine. Lots of amps and high voltage in these you going to have your guys open it up and mess around? Now you have the liability now your OEM can get pissed at you and not want to help because you messed with the design. You also compromised the UL label. The OEM can make a solid argument that if there is a workers comp suit that you altered it and hence it is on you. That's part of the reason why security screws are finding their way into factory equipment. Makes it that much harder to get inside.
I also question the mentality of customers who insist on individually long lasting parts when they haven't even established that they are. Sure NEMA contactors vs IEC ones look tougher but do you have data to back up that they are infact longer lasting? In real world conditions.
To put it bluntly: people get old and are afraid of change. They have half remembered wisdom about something that may have been true at one point but isn't anymore. The result is crap design. Fuses that don't protect anything, transformers that don't prevent surges, heafty parts that are a waste of material, line reactors when there is an internal one in the Soft starter/VFD, multiple voltage domains when 2 would have been sufficient, indicating lights that are dim, hardcoded non-safety interlocks when we have software, branch circuit protection when it isn't needed, expensive hard to replace push-to-test lights instead of a simple switch and software, thick wire crammed into limit switches that is almost impossible to repair, ancient PLCs no longer under support, serial comms that can't be troubleshooted, external power that defeats lock-out-tagout, roller switches that have to be made in a machine shop...
Oh sure your system will work but it radiants excessive heat, parts costs multiple times what they should to replace, large portions of the design do nothing but add heat and increase the chance of failure, it pulls a lot of energy, it can't be upgraded or even half the time incorporated into a line, it is more dangerous to work with, and you look like a fucking moron as people in the future learn not to buy a camel (horse built by committee) from you.
Don't go into controls/automation if you have any passion for good design. If you want to know what about a quarter of my customers are like imagine someone insisting that you apply the noose of a hangman to your forehead to fix a headache and women who show ankles carry smallpox.
That's all well and good for a factory, but I just want a dishwasher that lasts more than 5 years. I'm on my 3rd Bosch dishwasher. They keep getting worse. Now they require an app to use what was a button on the dishwasher.
I don't know. Maybe stop buying from them?
There are free solutions that are open source, hell there’s older commercial solutions from Acronis and whatever Symantec calls Ghost these days. You made a poor choice in selecting a losing horse in a race that’s been run many times - how is that a reflection of the state of modern tech? You didn’t choose the Hare, you chose poorly.
The article and this discussion isn’t about reliable solutions vs new fangled stuff that doesn’t realise, it’s about what we do now that stuff realised and we didn’t think about what we signed up for. I’m really glad your dad encouraged you to think about the value of well-worn approaches but you’re being extremely reductive as are many in this discussion. What I find interesting about that is I feel this trend towards reductive thinking probably reflects a world seemingly happily with sliding the Overton window right inch by inch.