Electronics > Projects, Designs, and Technical Stuff
Another reason to hate "soft touch" power switches
<< < (4/13) > >>
IDEngineer:

--- Quote from: SiliconWizard on July 18, 2019, 04:21:49 pm ---Anyway as we said: switched power strips. Done.
--- End quote ---
What about the supposed "can't let a hard power switch abruptly interrupt power or you'll corrupt internal data storage?" A power strip would risk that every single time.
IDEngineer:

--- Quote from: rsjsouza on July 18, 2019, 04:21:16 pm ---Power switches are critical for embedded OS equipment, as they can allow for the system to perform a proper shutdown. Clunking switches will leave temporary files behind and can cause filesystem corruption.
--- End quote ---
Then I guess a power strip isn't an option, either, despite many here recommending them as a solution.

The solution is to design equipment properly in the first place. There are plenty of products out there with transitory data that manage to not corrupt themselves through proper design. It's easy to know that power has been lost, have a predictable amount of uptime remaining due to latent energy in the power supply, and gracefully clean up before the rails actually die.
exe:

--- Quote from: IDEngineer on July 18, 2019, 04:51:42 pm ---
--- Quote from: SiliconWizard on July 18, 2019, 04:21:49 pm ---Anyway as we said: switched power strips. Done.
--- End quote ---
What about the supposed "can't abruptly interrupt power or you'll corrupt internal data storage?" A power strip would risk that every single time.

--- End quote ---

Just keep it read-only.
ejeffrey:
You can always use the soft-power button to initiate a shutdown and then use the power strip to cut power to the bench once everything is shut down.  But data corruption is mostly not a real problem since instruments have to handle power failures without bricking anyway.  Even desktop operating systems handle this pretty well these days and the embedded OS used in most instruments are even less likely to have data corruption problems on power loss.  Maybe that is an issue with high end instruments like VNAs or high speed oscilloscopes that cost more than a car or a house, but for your standard lab equipment this isn't a problem.
IDEngineer:

--- Quote from: exe on July 18, 2019, 04:55:59 pm ---
--- Quote from: IDEngineer on July 18, 2019, 04:51:42 pm ---
--- Quote from: SiliconWizard on July 18, 2019, 04:21:49 pm ---Anyway as we said: switched power strips. Done.
--- End quote ---
What about the supposed "can't abruptly interrupt power or you'll corrupt internal data storage?" A power strip would risk that every single time.
--- End quote ---
Just keep it read-only.
--- End quote ---
Sometimes not an option for "embedded OS equipment". However, proper design is an option, where you have a high priority interrupt triggered by loss of AC and you know the rails will survive for XXX milliseconds - during which your firmware's sole job is to stabilize its volatile data. Heck, even spinning disk drives have a version of this where they sense freefall and lock their heads into a safe position that doesn't risk damaging the platter surfaces. And that's in the mechanical realm.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod